Medical text conflict detection method, electronic device, and storage medium

Abstract

The application discloses a medical text conflict detection method, an electronic device and a computer readable storage medium. The method comprises the following steps: acquiring an electronic medical record text of a patient; performing conflict detection on the electronic medical record text based on preset conflict detection configuration information; and identifying conflict data in the electronic medical record text according to the result of the conflict detection. According to the application, the conflict data existing in the electronic medical record can be quickly and accurately found, and the influence on the patient's treatment is reduced.

Description

Technical Field

[0001] This application belongs to the field of text processing technology, and in particular relates to a medical text conflict detection method, electronic device, and computer-readable storage medium. Background Technology

[0002] Electronic medical records are records of all medical activities related to a patient and are linked to clinical reasoning. These records include various medical entities such as: factors, diseases, laboratory tests, treatments, medication regimens (both traditional Chinese and Western), health maintenance, and medical orders. These medical entities are interconnected and mutually constraining, creating a complex web of relationships.

[0003] Since electronic medical records are usually written manually by doctors, it is unavoidable that there will be conflicts between the medical entities recorded in the electronic medical records, which may affect the patient's medical treatment. Summary of the Invention

[0004] This application provides a medical text conflict detection method, electronic device, and computer-readable storage medium, which can quickly and accurately find conflicting data in electronic medical records and reduce the impact on patients' medical visits.

[0005] Firstly, this application provides a method for detecting conflicts in medical texts, including:

[0006] Obtain the patient's electronic medical record text;

[0007] Conflict detection is performed on the electronic medical record text based on the preset conflict detection configuration information;

[0008] Based on the results of conflict detection, conflicting data in the electronic medical record text is identified.

[0009] In a second aspect, this application provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the method described in the first aspect.

[0010] Thirdly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described in the first aspect above.

[0011] The advantages of this application compared to existing technologies are as follows: In this application, after obtaining the patient's electronic medical record text, the electronic device can perform conflict detection on the electronic medical record text based on preset conflict detection configuration information. Based on the conflict detection results, the electronic device can identify conflicting data in the electronic medical record text. This process does not rely on manual inspection of the electronic medical record by medical personnel, thus improving detection efficiency to a certain extent. Furthermore, users (such as medical personnel and / or medical institutions) can adjust the conflict detection configuration information according to their own habits or needs, which can further improve the accuracy and flexibility of this application in detecting conflicting data in electronic medical records. It is understood that the advantages of the second and third aspects mentioned above can be found in the relevant descriptions in the first aspect above, and will not be repeated here. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic diagram illustrating the implementation process of the medical text conflict detection method provided in the embodiments of this application;

[0014] Figure 2 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0015] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0016] To illustrate the technical solution proposed in this application, specific embodiments are described below.

[0017] The medical text conflict detection method proposed in this application is described below. The electronic device can be a personal computer, medical assistive device, smartphone, or server, or any other device with data processing capabilities. This application does not limit the specific type of the electronic device. Please refer to... Figure 1 The implementation process of this medical text conflict detection method is detailed below:

[0018] Step 101: Obtain the patient's electronic medical record text.

[0019] Electronic devices can obtain the patient's electronic medical record text through a string interface; alternatively, they can obtain the patient's electronic medical record text through file loading. This application embodiment does not limit the method of obtaining the electronic medical record text. It can be understood that the electronic medical record text obtained by the electronic device is specifically a large block of text content based on natural language.

[0020] Step 102: Perform conflict detection on the electronic medical record text based on the preset conflict detection configuration information.

[0021] After successfully acquiring a patient's electronic medical record text, the electronic device can initiate conflict detection on that text. This conflict detection relies on preset conflict detection configuration information. This configuration information is generally based on medical consensus literature and is used to detect conflicting medical data. Such medical consensus literature includes, but is not limited to, international / national medical standards, evidence-based guidelines, and clinical practice routines. As an example only, this medical consensus literature may include: *Clinical Diagnostics*, *Syndrome Differentiation*, and *International Standards for Traditional Medical Terminology*, etc., which will not be elaborated upon here.

[0022] Step 103: Based on the results of conflict detection, identify conflicting data in the electronic medical record text.

[0023] When data conflicts do exist in the electronic medical record text, the result of the conflict detection can indicate these data. These data are, in essence, conflicting data. Therefore, under the guidance of the conflict detection result, the electronic device can identify the conflicting data in the electronic medical record text. To facilitate user access and modification, the electronic device can centrally present the conflicting data in the electronic medical record text, specifically by marking the conflicting data. As an example only, the marking method can be based on highlighted color, or on font weight and / or font size; this application embodiment does not limit the marking method used by the electronic device.

[0024] In some embodiments, since electronic medical record texts contain various aspects of patient-related information, this embodiment optimizes the medical text conflict detection method to achieve ordered conflict detection of electronic medical record texts and improve the efficiency of conflict detection. The optimized medical text conflict detection method includes:

[0025] Step 201: Obtain the patient's electronic medical record text.

[0026] It is understood that step 201 is the same as or similar to step 101. For details, please refer to the previous description of step 101, which will not be repeated here.

[0027] Step 202: Segment the electronic medical record text to obtain the medical record paragraph text corresponding to each segment title in the electronic medical record text.

[0028] Electronic devices can parse electronic medical record text using Natural Language Processing (NLP) technology. Specifically, this involves segmenting the electronic medical record text using preset keywords. After segmentation, the electronic device obtains multiple segment headings, either explicitly or implicitly contained within the electronic medical record text. Based on this, the text between the i-th segment heading and the (i+1)-th segment heading is the medical record paragraph text corresponding to that i-th segment heading, and the text from the N-th segment heading up to the end of the electronic medical record text is the medical record paragraph text corresponding to that N-th segment heading. Here, i is a positive integer greater than 0 and less than N, and N is the total number of segment headings contained in the electronic medical record text. Thus, the medical record paragraph text corresponding to each segment heading can be obtained.

[0029] In some examples, to facilitate subsequent detection and processing, the medical record paragraph text corresponding to each segment title can be represented in the form of a list of short medical record text sentences; of course, the medical record paragraph text corresponding to each segment title can also be represented in other forms, which are not limited here.

[0030] Step 203: Extract the entity list corresponding to at least one medical entity type from the electronic medical record text.

[0031] Electronic devices can use NLP technology to parse electronic medical record text, specifically to extract a list of entities corresponding to at least one type of medical entity.

[0032] In some examples, the entity lists corresponding to at least one medical entity type include, but are not limited to: an element list, a test protocol list, an examination protocol list, a list of Western medicines, and a disease list; the element list also includes: a positive element list and a negative element list. The element list stores all elements contained or implicitly contained in the electronic medical record text. Elements refer to disease-related characteristics possessed by the patient; negative elements refer to elements used to represent the patient's positive health status; and positive elements refer to elements used to represent the patient's negative health status.

[0033] Step 204: Detect conflicts based on preset conflict detection configuration information, using each medical record paragraph text as the detection object and / or each entity list as the detection object.

[0034] It is understandable that the medical record paragraph text and entity list can be considered as products obtained by disassembling the electronic medical record text. When performing conflict detection, electronic devices can use the medical record paragraph text and / or entity list as the detection objects to achieve conflict detection based on preset conflict detection configuration information.

[0035] Step 205: Based on the results of conflict detection, identify conflicting data in the electronic medical record text.

[0036] It is understood that step 205 is the same as or similar to step 103. For details, please refer to the previous description of step 103, which will not be repeated here.

[0037] In some embodiments, manually written electronic medical records may not only have data conflicts but also data non-standardization issues. To address the data non-standardization problem, this embodiment optimizes the medical text conflict detection method described above. The implementation process of this optimized medical text conflict detection method is detailed below:

[0038] Step 301: Obtain the patient's electronic medical record text.

[0039] Step 302: Segment the electronic medical record text to obtain the medical record paragraph text corresponding to each segment title in the electronic medical record text.

[0040] Step 303: Extract the entity list corresponding to at least one medical entity type from the electronic medical record text.

[0041] Step 304: Detect conflicts based on preset conflict detection configuration information, using each medical record paragraph text as the detection object and / or each entity list as the detection object.

[0042] Step 305: Based on the results of conflict detection, identify conflicting data in the electronic medical record text.

[0043] It is understood that steps 301-305 are the same as or similar to steps 201-205, respectively. For details, please refer to the previous description of steps 201-205, which will not be repeated here.

[0044] Step 306: Based on the preset standardization detection configuration information, perform standardization detection on each segment title and each medical record paragraph text.

[0045] After successfully acquiring the patient's segment headings and the corresponding medical record paragraph text, the electronic device can initiate a standardization check on those headings and paragraph text. This standardization check relies on preset standardization check configuration information. Similar to conflict detection configuration information, this standardization check configuration information is generally configured based on medical consensus literature to detect non-standard medical data, and will not be elaborated here.

[0046] Step 307: Based on the results of the standardization test, identify non-standard data in the electronic medical record text.

[0047] If non-standard data is indeed found in the electronic medical record text, the result of this standardization check can indicate the nature of this data. This data can be understood as non-standard data. Therefore, guided by the result of this standardization check, the electronic device can identify the non-standard data in the electronic medical record text. To facilitate user review and modification, the electronic device can centrally present conflicting data in the electronic medical record text, specifically by marking the non-standard data. The method used to mark non-standard data is similar to the method used to mark conflicting data described earlier, and will not be repeated here.

[0048] In some embodiments, conflict detection configuration information includes, but is not limited to, a main conflict table and sub-conflict tables, as well as warning scoring rules, etc., which are not limited here. For ease of understanding, each type of conflict detection configuration information is described and introduced in detail below:

[0049] For conflicting main tables and conflicting sub-tables:

[0050] The conflict master table is configured with multiple conflict master data entries. The terms for each conflict master data entry can be represented as: {Id|ClashCount|ShowType|MarkContent|SuggestedContent|Other}.

[0051] The conflict sub-table is configured with conflict sub-data corresponding to each conflicting master data entry. The terms for each conflict sub-data entry can be represented as follows:

[0052] {Id|MasterId|Type|TblName|TblId|TblCName|MinValue|MaxValue|UnitValue|Other}.

[0053] For terms related to conflicting master data: Id is the conflicting master data ID, used to uniquely identify a conflicting master data, corresponding to the MasterId in the conflicting sub-table; ClashCount indicates the number of conflicting sub-data entries corresponding to the conflicting master data; ShowType indicates the interface display type of the conflicting master data, defined as: 1 for conditional warning, 2 for contradictory mutual exclusion; MarkContent indicates the marked content of the conflicting master data; SuggestedContent indicates the suggested content of the conflicting master data; Other indicates other information about the conflicting master data, such as the user and editing time.

[0054] For terms in conflicting subdata: Id is the Id of the conflicting subdata itself, used to uniquely identify a conflicting subdata; MasterId represents the Id of the conflicting master data corresponding to the conflicting subdata, which corresponds to the Id in the conflicting master table; Type represents the term type of the conflicting subdata, defined as: 0 for other, 1 for sex, 2 for age, 3 for sym, 4 for disease, 5 for Western medicine, 6 for laboratory, 7 for detection, and 8 for traditional Chinese medicine; TblName represents the name of the entity list corresponding to the term type of the conflicting subdata, such as sym or ill; TblId represents the Id of the entity in the corresponding medical knowledge base (i.e., entity list) of the conflicting subdata; TblCName represents the name of the conflicting subdata; MinValue, MaxValue, and UnitValue all represent the thresholds corresponding to conflict detection when numerical judgment is required, where MinValue represents the minimum value, MaxValue represents the maximum value, and UnitValue represents the unit; Other represents other information of the conflicting subdata, such as the user and editing time.

[0055] It can be understood that a conflict master data and all the conflict sub-data corresponding to that conflict master data constitute a complete conflict detection rule.

[0056] It should be noted that when the conflicting master table and conflicting child table are associated and mapped through the master table ID and the child table MasterId, the following conditions must be met simultaneously:

[0057] 1) The number of conflicting sub-data with the same MasterId in the conflicting sub-table should be equal to the ClashCount recorded in the conflicting master data with Id=MasterId in the conflicting master table.

[0058] The following gives relevant examples: In the conflict master table, if the ClashCount recorded for the conflict master data with Id 9502 is 2, then in the conflict sub-table, there should be exactly 2 conflict sub-data records with MasterId 9502. If there are more or fewer than 2 conflict sub-data records with MasterId 9502, it indicates that the configuration of the conflict master table and the conflict sub-table is incorrect, and a reminder message can be output to prompt the user to adjust the conflict sub-table and / or the conflict master table.

[0059] 2) In the conflict sub-table, if Type is 3 to 8, then the corresponding TblName and TblId need to be configured simultaneously; if Type is 0 to 2, then there is no need to configure TblName and TblId. The reason is as follows: When Type is 3 to 8, it indicates that the corresponding entry is a standard medical entity entry, so TblName and TblId need to be configured; when Type is 0, it indicates that the corresponding entry may be a non-standard medical entity entry, when Type is 1, it indicates that the entry type of the corresponding entry is gender (that is, this entry is used to identify the gender of the patient), when Type is 2, it indicates that the entry type of the corresponding entry is age (that is, this entry is used to identify the age of the patient). In these three cases, there is no need to configure TblName and TblId. It can be understood that when the conflict master table and the conflict sub-table are edited by the user, they still need to meet the above requirements.

[0060] For the early warning scoring rules:

[0061] The early warning scoring rules can be subdivided into the MEWS (modified early warning score, MEWS) early care warning scoring rules and the PEWS (pediatric early warning score, PEWS) early care warning scoring rules. Among them, the target of the MEWS early care warning scoring rules is adults, and the target of the PEWS early care warning scoring rules is children.

[0062] Each entry of the MEWS early care warning scoring rules can be expressed as:

[0063] <MEWSRules|{Item1|Score=0,R|Score=1,R|Score=2,R|Score=3,R},{Item2|Score=0,R|Score=1,R|Score=2,R|Score=3,R},…,{Disposal suggestions|SumScore1, Patient condition 1, Treatment measure 1|SumScore2, Patient condition 2, Treatment measure 2|…}>

[0064] For the entry of the MEWS early care warning score rules:

[0065] Item1, Item2, etc. are the key items to be monitored for early care of adults, and the number of key items is equal to the number of configured Items. Only as an example, the key items include body temperature, respiration, heart rate, etc., which are not limited here.

[0066] Score represents the score, indicating the scoring level, and can take values between 0 and 3. Of course, some scoring levels can also be left blank, which means that the content corresponding to this scoring level can be not configured. Only as an example, the highest Score for body temperature is 2, so the content with Score of 3 can not be configured in the entry.

[0067] R represents the monitoring value range of the corresponding key item under different scores. SumScore represents the accumulation of scores for each key item. Only as an example, if the Score of Item1 is 1, the Score of Item2 is 3, and the Score of other Item items is 0, then SumScore is 4.

[0068] The disposal suggestions represent the situation descriptions and treatment measure suggestions given to patients respectively according to different scores of SumScore.

[0069] In some examples, taking the body temperature monitoring item as an example, if the corresponding MEWS early care warning score rule is: <MEWS Rules|{body temperature (°C)|Score = 0, R = (36.1, 38)|Score = 1, R = (35.1, 36) OR R = (38.1, 38.5)|Score = 2, R <= 35.0 OR R >= 38.6}……>, it means that when the patient's body temperature is between 36.1°C and 38°C, Score is 0; when the body temperature is between 35.1°C and 36°C or between 38.1°C and 38.5°C, Score is 1; when the body temperature is less than or equal to 35.0°C or greater than or equal to 38.6°C, Score is 2; since the highest Score for the key item of body temperature is 2, the situation with Score of 3 is not configured in this rule.

[0070] Each entry of the PEWS early care warning score rules can be expressed as:

[0071] <PEWSRules|{Item1|Score = 0,R|Score = 1,R|Score = 2,R|Score = 3,R},{Item2|Score = 0,R|Score = 1,R|Score = 2,R|Score = 3,R},…,{Disposal Suggestion|SumScore1 / Score1,Handling Measure|SumScore2 / Score2,Patient Condition,Handling Measure|…}>.

[0072] For the entries of the PEWS early care warning score rules, the meanings of each parameter are the same as those in the MEWS early care warning score rules, and will not be elaborated here.

[0073] However, it should be noted that in the PEWS early care warning score rules, in addition to giving disposal suggestions based on the value of SumScore, disposal suggestions can also be given based on the value of individual Score. In some examples, if the corresponding PEWS early care warning score rules are: <PEWS Rules|……,{Disposal Suggestion|…|SumScore> = 4 OR Score = 3,Patient Condition ④,Handling Measure ④}>, it means that if SumScore is greater than or equal to 4, or if the Score of any one of several key items for early care monitoring of children reaches 3, the suggestion of Handling Measure ④ can be initiated.

[0074] Users can modify the monitored value ranges corresponding to each score in the MEWS and PEWS early care warning score rules according to their own needs (that is, they can modify the R corresponding to each Score), and can also modify the disposal suggestions, but generally do not allow adding or deleting the corresponding key items and scores (that is, do not modify Item and Score).

[0075] Of course, according to the requirements of the actual usage scenario, the conflict detection configuration information can also include: other conflict detection rules. Only as an example, the other conflict detection rules can be surgical anesthesia risk prompt detection rules, drug sensitivity analysis detection rules, and detection rules for duplicate items in the positive element list and negative element list, etc., and the embodiments of this application do not limit this.

[0076] In some embodiments, the normative detection configuration information includes but is not limited to: basic information normative detection rules, section title integrity detection rules, and medical record paragraph text content normative detection rules, etc., which are not limited here. For ease of understanding, each type of normative detection configuration information will be specifically described and introduced below:

[0077] For the basic information normative detection rules;

[0078] The basic information standardization detection rules are mainly set for the following three key basic information: name, age, and gender. The entries of the basic information standardization detection rules can be expressed as:

[0079] <Key Factors|{Name|NotEmpty = 1, BText = "Name: "}, {Age|NotEmpty = 1, BText = "Age: ", AgeUnit1 = "years", Range1 = (0, 120), AgeUnit2 = "months", Range2 = (1, 12), AgeUnit3 = "days", Range3 = (1, 31)}, {Sex|NotEmpty = 1, BText = "Gender: "}>

[0080] Among them, Name is used to mark the name, that is, the following is the standardization detection rule for the name; NotEmpty = 1 means that the name of the patient in the electronic medical record text cannot be empty (of course, if NotEmpty = 0, it means that the name can be empty); BText = "Name: " means that the name of the patient needs to be after the text "Name: ".

[0081] Age is used to mark the age, that is, the following is the standardization detection rule for the age; NotEmpty = 1 means that the age of the patient in the electronic medical record text cannot be empty (of course, if NotEmpty = 0, it means that the age can be empty); BText = "Age: " means that the age of the patient needs to be after the text "Age: "; AgeUnit1, AgeUnit2, and AgeUnit3 indicate that there are 3 units for age. Among them, the age unit 1 is "years", and its range is 0 - 120; the age unit 2 is "months", and its range is 1 - 12; the age unit 3 is "days", and its range is 1 - 31.

[0082] Sex is used to mark the gender, that is, the following is the standardization detection rule for the gender; NotEmpty = 1 means that the gender of the patient in the electronic medical record text cannot be empty; BText = "Gender: " means that the gender of the patient in the electronic medical record text needs to be after the text "Gender: ".

[0083] Users can set the value of NotEmpty for each basic information and the specific content of BText in the basic information standardization detection rules according to their needs, and can also set the value of Range for each age unit of age. There is no limitation here.

[0084] For the detection rules of the integrity of the segmented title:

[0085] The segment title integrity check rule is used to check whether the segment titles contained in the electronic medical record text are complete. The terms of this segment title integrity check rule can be represented as: <title norms,list>。< / title>

[0086] Here, `list` represents the section headings that must be included in the electronic medical record text. As an example only, if the section heading integrity check rule is: <title norms,{"主诉”,”现病史”,”既往史”,”辅助检查”,”体格检查”,”诊断”}> This indicates that the electronic medical record text must include the following section headings: "Chief Complaint," "Present Illness," "Past Medical History," "Auxiliary Examinations," "Physical Examination," and "Diagnosis," along with their corresponding medical record paragraph text. It's understandable that if the electronic medical record text contains missing section headings, empty corresponding paragraph texts, or duplicate section headings, this section heading integrity check rule can detect these irregularities and provide appropriate prompts.< / title>

[0087] Users can add, delete, or modify the segment titles to be detected as needed; there are no restrictions here.

[0088] Regarding the rules for checking the standardization of medical record text content:

[0089] The rules for detecting the standardization of content in medical record paragraphs are used to check the standardization of the content contained in medical record paragraphs. The terms of these rules can be represented as follows:<string,list> .

[0090] Wherein, `string` represents the section headings corresponding to the medical record paragraphs that require content standardization checks, and `list` represents the content standardization check rules for those medical record paragraphs. It should be noted that not all medical record paragraphs corresponding to section headings require content standardization checks. In this embodiment, for ease of explanation, we take the content standardization check of the medical record paragraphs corresponding to the chief complaint section heading as an example, and provide an example of a content standardization check rule for medical record paragraphs, which can be:

[0091] <Chief complaint, {Length<21|Text Limit=disease, negation word, “several days”}>.

[0092] For the medical record paragraph corresponding to the chief complaint, the content standardization check rule is: Length < 21 | Text Limit = Disease, Negative Word, "Several Days". Here, Length < 21 means the text length of this medical record paragraph must be less than 21 characters (because the chief complaint in electronic medical records needs to be concise and clear); Text Limit indicates the specific content restriction for this medical record paragraph. In this example, Text Limit = Disease, Negative Word, "Several Days" means that the chief complaint cannot contain disease names, negative words, or vague descriptive language such as "several days". Disease names can be obtained from a disease thesaurus, and negative words can be obtained from a negative word dictionary.

[0093] Users can set limits on the length as needed, and can also add, delete, or modify the specific content of the TextLimit; no restrictions are imposed here. The modifiable content of the TextLimit generally includes vague text that should not be included in the main complaint, such as "several days."

[0094] In some embodiments, various detection configuration information (including but not limited to conflict detection configuration information and normative detection configuration information), a disease thesaurus, and a negation word dictionary may be stored in a medical knowledge base. As an example only, this medical knowledge base may be deployed in the cloud, and the electronic device may load the required detection configuration information from the cloud when needed; alternatively, the medical knowledge base may also be deployed locally on the electronic device, and the required detection configuration information may be loaded from that local device when needed; this is not limited here.

[0095] The various testing configuration information stored in the medical knowledge base can be adjusted according to the needs or habits of users (such as medical personnel and / or medical institutions). Specifically, the medical knowledge base provides users with a management interface. Through this management interface, users can query and edit the content stored in the medical knowledge base. Editing operations include adding, deleting, and / or modifying, thereby ensuring that the medical knowledge base can be continuously improved in practical applications.

[0096] In some embodiments, to ensure the data security of the medical knowledge base, the management interface also has a management retrospective function; that is, all editing operations performed by the user through the management interface, the time corresponding to each editing operation, and the user information corresponding to the user can be recorded by the medical knowledge base. Specifically, the deletion operation marks the corresponding content as deleted, rather than directly removing the corresponding content from the medical knowledge base.

[0097] In some embodiments, the management permissions corresponding to different detection configuration information are not the same. For example, the conflict master table and conflict sub-table in the conflict detection configuration information support adding, deleting, modifying and querying operations; however, the prescriptive detection configuration information and the early warning scoring rules and other conflict detection rules in the sudden detection configuration information only support querying and partial modification operations, and do not support adding and deleting operations.

[0098] It is understandable that the content stored in the medical knowledge base can be continuously optimized through this user-facing management interface, thereby continuously improving the clinical value of the medical text conflict detection method proposed in this application embodiment.

[0099] In some embodiments, the medical knowledge base is physically stored in a database. Therefore, to ensure data consistency, the electronic device can employ the database's transaction mechanism when performing add, delete, and modify operations; that is, after initiating a transaction, the corresponding operation is performed. If the operation is normal, the transaction is committed uniformly; if the operation is abnormal, the transaction is rolled back.

[0100] In some embodiments, the conflict detection operation described above, which uses each medical record paragraph text as the detection object and / or each entity list as the detection object and is based on preset conflict detection configuration information, may specifically include:

[0101] A1. Obtain the patient's target basic information from the electronic medical record text.

[0102] As described above, the patient's basic information mainly includes the patient's name, age, and gender. Of these three types, the patient's age and gender are more closely related to the patient's various disease manifestations and characteristics, and thus belong to basic information from an objective perspective. The patient's name, on the other hand, is largely unrelated to the patient's various disease manifestations and characteristics, and belongs to basic information from a subjective perspective. Therefore, this embodiment uses the patient's age and gender as the target basic information. Through NLP processing and semantic analysis, the electronic device can obtain the patient's target basic information from the patient's electronic text.

[0103] A2. Clean the conflict detection configuration information based on the target's basic information to obtain the target conflict detection configuration information.

[0104] Some conflict detection configuration information has limitations in its applicability. For example, some conflict detection configuration information is only applicable to a specific gender, and others are only applicable to a specific age group. Therefore, not all conflict detection configuration information is necessary for the current patient's electronic medical record text. Based on this, the electronic device can perform data cleaning based on the patient's target basic information to filter out the necessary conflict detection configuration information, i.e., the target conflict detection configuration information.

[0105] It is important to note that the cleaning operation performed here targets the already loaded conflict detection configuration information. This cleaning operation will not affect the medical knowledge base; that is, the conflict detection configuration information stored in the medical knowledge base remains complete and uncleaned.

[0106] In some embodiments, the conflict master table and conflict sub-tables have been described above and will not be repeated here. As described above, some entries in the conflict master table and conflict sub-tables are also related to specific entity lists. Therefore, when cleaning the conflict master table and conflict sub-tables, it is necessary to consider not only the target basic information but also the entity lists extracted from the electronic medical record text. This step can be specifically manifested as: cleaning the conflict detection configuration information according to the target basic information and the medical entity type to which each entity list belongs, to obtain the target conflict detection configuration information.

[0107] Specifically, the process of cleaning the conflict master table and conflict sub-tables based on the patient's gender is as follows: For conflict sub-data with type 1 in the conflict sub-table, the value of TblCName in the conflict sub-data is obtained and matched with the patient's gender; if there is no match, it means that the conflict sub-data does not need to be checked. The electronic device can use the MasterId in the conflict sub-data to map and find the conflict master data in the conflict master table, and delete the conflict master data and conflict sub-data with the same MasterId in the conflict sub-table. After traversing all conflict sub-data with type 1 in the conflict sub-table, the cleaning of the conflict master table and conflict sub-table based on the patient's gender is completed.

[0108] Specifically, the process of cleaning the conflict master table and conflict sub-tables based on the patient's age is as follows: For conflict sub-data of type 2 in the conflict sub-table, obtain the MinValue and MaxValue values ​​of the conflict sub-data and compare them with the patient's age. If the comparison result satisfies either of the following two conditions: 1) The conflict sub-table has a configured MinValue value, and the patient's age is less than the MinValue value; 2) The conflict sub-table has a configured MaxValue value, and the patient's age is greater than the MaxValue value, then it means that the conflict sub-data does not match the current patient, and the conflict sub-data does not need to be detected. The electronic device can use the MasterId in the conflict sub-data to map and find the conflict master data in the conflict master table, and delete the conflict master data and conflict sub-data with the same MasterId in the conflict sub-table. After traversing all conflict sub-data of type 1 in the conflict sub-table, the cleaning of the conflict master table and conflict sub-table based on the patient's gender is completed.

[0109] It is understandable that if the electronic device cannot obtain the patient's gender from the electronic medical record text, then the electronic device does not need to perform the operation of cleaning the conflict main table and conflict sub-table based on the patient's gender; similarly, if the electronic device cannot obtain the patient's age from the electronic medical record text, then the electronic device does not need to perform the operation of cleaning the conflict main table and conflict sub-table based on the patient's age.

[0110] Specifically, the process of cleaning the conflict master table and conflict sub-tables based on the entity list is as follows: The entity list contains multiple entities, each with its own ID. Taking the element list as an example, this element list contains multiple elements, each with its own ID, used to uniquely identify the element. Therefore, the electronic device can clean the conflict master table and conflict sub-tables based on the IDs of each entity extracted from the entity lists, deleting the corresponding conflicting master data in the conflict master table and the corresponding conflicting sub-data in the conflict sub-tables. Taking the element list as an example, the implementation process is as follows: In the conflict sub-tables, look for conflicting sub-data with TblName "sym" and whose TblId does not contain the IDs of any element in the element list. These conflicting sub-data are considered not to match the current patient's electronic medical record text and do not need to be detected. The TblId of these conflicting sub-data is mapped to the MasterId, and then the conflicting master data in the conflict master table can be mapped and found through this MasterId. The conflicting master data and conflicting sub-data with the same MasterId in the conflict sub-tables are then deleted.

[0111] It is understandable that the process of cleaning the conflict master table and conflict sub-tables based on the test protocol list, examination protocol list, Western medicine list, and disease list is basically the same as the process of cleaning the conflict master table and conflict sub-tables based on the entity list, except that the corresponding TblName values ​​are different. Specifically, the TblName corresponding to the test protocol list is "lab", the TblName corresponding to the examination protocol list is "chk", the TblName corresponding to the Western medicine list is "dru", and the TblName corresponding to the disease list is "ill". This will not be elaborated further here.

[0112] In some embodiments, the warning scoring rules have been described above and will not be repeated here. As described above, the warning scoring rules are subdivided into MEWS early care warning scoring rules for adults and PEWS early care warning scoring rules for children. Therefore, when cleaning the warning scoring rules, the patient's age is the primary consideration. It can be understood that if the patient is older than 14 years old, the electronic device can clean the PEWS early care warning scoring rules and retain the MEWS early care warning scoring rules as target conflict detection configuration information; conversely, if the patient is younger than or equal to 14 years old, the electronic device can clean the MEWS early care warning scoring rules and retain the PEWS early care warning scoring rules as target conflict detection configuration information.

[0113] In some embodiments, other conflict detection rules, as described above, will not be repeated here. As described above, these other conflict detection rules have low correlation with the patient's age, gender, and entity list; therefore, the electronic device does not need to clean these other conflict detection rules. That is, the electronic device can treat all other conflict detection rules as target conflict detection configuration information.

[0114] A3. Detect conflict based on the target conflict detection configuration information, using each medical record paragraph text as the detection object and / or each entity list as the detection object.

[0115] When electronic devices use the text of each medical record segment as the detection object and / or use each entity list as the detection object, they no longer need to perform conflict detection based on all the conflict detection configuration information originally stored in the medical knowledge base, but only need to perform conflict detection based on the target conflict detection configuration information that has been filtered.

[0116] In some embodiments, for ease of explanation, the remaining conflict master table and conflict sub-table after cleaning are referred to as the target conflict master table and target conflict sub-table; that is, for the conflict master table and conflict sub-table, the target conflict detection configuration information includes: the target conflict master table and the target conflict sub-table. Based on the target conflict master table and target conflict sub-table, the conflict detection process can be specifically as follows: traversing the target conflict master table; obtaining the target conflict sub-table corresponding to the currently traversed conflict master data (that is, obtaining the conflict sub-data corresponding to the currently traversed conflict master data in the target conflict sub-table, which can be referred to as the target conflict sub-data); matching the target conflict sub-table with each entity list and / or each medical record paragraph text (that is, matching each target conflict sub-data with each entity list and / or each medical record paragraph text) to determine whether a conflict exists. The process is described in detail below:

[0117] B1. Traverse the target conflict master table. Since the MasterId of the conflicting sub-data is the same as the Id of the corresponding conflicting master data, the corresponding conflicting sub-data can be found based on the Id in the currently traversed conflicting master data. The electronic device can write the found conflicting sub-data into a temporary sub-table rule list and compare the number of records in the temporary sub-table rule list with the ClashCount value recorded in the conflicting master data corresponding to that Id in the target conflict master table. If they are not equal, it indicates that the group of conflicting sub-data and / or the corresponding conflicting master data is abnormal, and no detection is needed; the system can return to the target conflict master table to traverse the next conflicting master data. If they are equal, proceed to the next step.

[0118] B2. Define a temporary variable `tempClashTips` for conflict alerts. Electronic devices can iterate through the temporary sub-table rule list. Different operations can be performed based on the Type or TblName of the conflicting sub-data encountered during iteration, specifically:

[0119] 1) If the Type of the conflicting sub-data encountered during the iteration is 1 (representing gender) or 2 (representing age), then add the conflicting sub-data to tempClashTips.ClashList.

[0120] 2) If the Type of the conflicting sub-data encountered during the iteration is 3 (representing a feature) or the TblName is "sym", then search the extracted feature list for data with an Id equal to the TblId of the conflicting sub-data. If such data is found, add the conflicting sub-data and the found data to tempClashTips.ClashList. If such data is not found, proceed to step 6) for further processing.

[0121] 3) If the Type of the conflicting sub-data is 4 (representing a disease) or the TblName is "ill", then search the extracted disease list for data with an Id equal to the TblId of the conflicting sub-data. If no such data is found, proceed to step 6). If such data is found, continue to determine whether the conflicting sub-data has configured MinValue and MaxValue values. If neither MinValue nor MaxValue is configured, add the conflicting sub-data and the found data to tempClashTips.ClashList. If MinValue and / or MaxValue are configured, proceed to step 6).

[0122] 4) Similarly, if the Type of the conflicting sub-data encountered is 5 (representing Western medicine), 6 (representing inspection), or 7 (representing check), then refer to the implementation process given in step 3), only the corresponding TblName is different, which will not be repeated here.

[0123] 5) If the Type of the conflicting sub-data encountered during the traversal is 0 (indicating other types), then directly jump to step 6) for processing.

[0124] 6) Search for the content of TblCName in the conflicting sub-data within each medical record paragraph (or, alternatively, within the electronic medical record text). If no data is found, end this step directly. If data is found, write the medical record content including TblCName into tempClashTips.ClashList. Specifically, this can be divided into the following two cases:

[0125] Scenario 1: The MinValue and MaxValue of the conflicting sub-data being iterated over are not configured. In this case, the electronic device can add the conflicting sub-data and the found medical record content to tempClashTips.ClashList, and continue iterating over the next conflicting sub-data in the temporary sub-table rule list.

[0126] The second scenario: The encountered conflicting sub-data contains configured MinValue and / or MaxValue values. In this case, the electronic device can extract the string between the corresponding TblCName text in the medical record content and the UnitValue text configured in the conflicting sub-data. It then uses a regular expression to obtain the numerical value from this string and matches it against the configured MinValue and / or MaxValue in the conflicting sub-data. If a match is successful, the conflicting sub-data and the found medical record content are added to tempClashTips.ClashList, and the system continues to iterate through the next conflicting sub-data in the temporary sub-table rule list. The conditions for a successful match are: if both MinValue and MaxValue have values, the numerical value obtained through the regular expression must be between MinValue and MaxValue; if MinValue has a value but MaxValue has no value, the numerical value obtained through the regular expression must be greater than MinValue; if MinValue has no value but MaxValue has a value, the numerical value obtained through the regular expression must be less than MaxValue.

[0127] B3. After the electronic device has finished traversing the temporary sub-table rule list, if the total number of conflicting sub-data in the temporary sub-table rule list is greater than the total number of contents stored in tempClashTips.ClashList, it means that there is no match; otherwise, it means that there is a match, and the electronic device can add the temporary variable tempClashTips to the conflict warning list.

[0128] B4. Once the electronic device has completed its traversal of the target conflict master table, it can be determined that the conflict detection based on the target conflict master table and the target conflict sub-table has been completed.

[0129] In some embodiments, for ease of explanation, the warning scoring rules retained after cleaning are referred to as the target warning scoring rules; that is, for the warning scoring rules, the target conflict detection configuration information includes: the target warning scoring rules. Based on the warning scoring rules, the conflict detection process can be specifically as follows: traversing the target warning scoring rules; matching the currently traversed key items with the text of each medical record segment; determining the patient's score on the currently traversed key items based on the matching results and the scores corresponding to the numerical range of the key items; and accumulating the patient's scores on each key item to determine the result of the conflict detection. The following explanation uses the MEWS early care warning scoring rule and the PEWS early care warning scoring rule as examples to illustrate this process:

[0130] When the target early warning scoring rule is the MEWS early care early warning scoring rule:

[0131] C1. When the key item encountered during the traversal is body temperature, the electronic device can obtain the patient's body temperature data from the medical record text using NLP technology. Specifically, this can be divided into the following three situations:

[0132] In the first scenario: if the electronic device does not obtain the patient's body temperature data from the medical record text, the score for this key item, body temperature, can be determined as 0.

[0133] The second scenario: The electronic device obtains a patient's temperature data from the medical record text. Following the Score ranking from highest to lowest, the electronic device compares this temperature data with the monitoring ranges for different Score scores in the MEWS Early Care Warning Scoring Rules. Once a monitoring range is matched, the electronic device can end the comparison and determine the score for the key item—temperature—as the Score corresponding to that monitoring range.

[0134] The third scenario: The electronic device obtains more than two temperature readings from the patient's medical record text. For each temperature reading, the electronic device can process it according to the procedure described in the second scenario, thereby obtaining multiple temperature scores. The electronic device can then select the highest score as the score for the key item of temperature.

[0135] For example, in the electronic medical record text of an electronic device, the body temperature data of the patient obtained through NLP parsing is 38.2°C. When comparing it with the range of body temperature in the MEWS early care warning scoring rule, the monitoring value range (R <= 35.0 OR R >= 38.6) with Score = 2 can be compared with this body temperature data first, and the result is a miss. Due to the miss, the electronic device can then compare the monitoring value range (35.1 < R < 36 OR 38.1 < R < 38.5) with Score = 1 with this body temperature data, and the result is a hit. Finally, the electronic device can determine that the score for the key item of body temperature is 1.

[0136] C2. When the key items traversed are respiration, heart rate, systolic blood pressure, blood glucose, blood oxygen saturation, urination, etc., the implementation mechanism is the same as that for body temperature, and will not be elaborated here.

[0137] C3. When the key item traversed is the level of consciousness, the electronic device can sequentially search for the specified text content in the MEWS early care warning scoring rule in the medical record paragraph text in the order of decreasing Score. As long as one specified text content is found in the medical record paragraph text, the electronic device can use the Score corresponding to the specified text content as the score for the key item of the level of consciousness and end the processing. If no specified text content is found, the score for the key item of the level of consciousness is 0.

[0138] C4. After the electronic device finishes traversing the key items, the scores of each key item are accumulated to obtain the total score.

[0139] C5. The electronic device can compare the calculated total score with the value shown in the treatment suggestions in the MEWS early care warning scoring rule to determine the treatment measures for the current patient and write them into the conflict prompt list.

[0140] In the case where the target warning scoring rule is the PEWS early care warning scoring rule:

[0141] D1. When the key item traversed is consciousness, the electronic device can sequentially search for the specified text content shown for the key item of consciousness in the PEWS early care warning scoring rule in the medical record paragraph text in the order of decreasing Score. As long as one specified text content is found in the medical record paragraph text, the electronic device can use the Score corresponding to the specified text content as the score for the key item of consciousness and end the processing. If no specified text content is found, the score for the key item of the degree of consciousness is 0.

[0142] D2. When the key item encountered during the iteration is heart rate, the PEWS early care warning scoring rules include both text content matching and comparison based on heart rate value ranges. Therefore, the electronic device needs to determine the score according to both of these rules:

[0143] First, a text content matching comparison is performed. The electronic device can match the specified text content corresponding to the highest score in the medical record text, from highest to lowest score. After a specified text content is matched, the score corresponding to that text content is temporarily taken as the heart rate score, and the text content matching comparison process ends. If no match is found, the heart rate score is temporarily recorded as 0.

[0144] Then, the heart rate values ​​are compared within a range. After the electronic device obtains the patient's heart rate data from the medical record text using NLP technology, there are three possibilities:

[0145] The first scenario: When the electronic device does not obtain the patient's heart rate data from the medical record text. In this case, the electronic device directly uses the score obtained by matching the text content as the final score for the key item of heart rate.

[0146] The second scenario: The electronic device obtains a patient's heart rate data from a medical record text segment. In this case, the heart rate data is compared with the monitoring value range corresponding to each score, arranged in descending order of score. After matching any monitoring value range, the score corresponding to that monitoring value range is compared with the score obtained by matching the text content, and the larger score is taken as the final score for the key item of heart rate.

[0147] The third scenario: The electronic device obtains multiple heart rate data points from the patient's medical record text. In this case, for each heart rate data point, the implementation mechanism in the second scenario is followed to obtain multiple scores corresponding to the heart rate data points. The largest score is then compared with the score obtained by matching the text content, and the larger score is taken as the final score for the key item of heart rate.

[0148] D3. When iterating through other key items, if the PEWS early care warning scoring rules also contain two rules: text content matching and comparison based on the range of respiratory values, then the implementation mechanism is the same as that of the respiratory key item, which will not be elaborated here.

[0149] D4. After the electronic device has completed its traversal of the key items, the scores of each key item are summed to obtain the total score.

[0150] D5. The electronic device can compare the calculated total score and / or the scores of each key item with the values ​​shown in the treatment recommendations in the PEWS Early Care Warning Scoring Rules to determine the appropriate treatment measures for the current patient and write them into the conflict alert list.

[0151] In some embodiments, the normative detection configuration information described above includes, but is not limited to: basic information normative detection rules, segment title integrity detection rules, and medical record paragraph text content normative detection rules. The following is a detailed explanation of the normative detection process based on these various normative detection configuration information:

[0152] For normative testing based on basic information normative testing rules:

[0153] E1. Electronic devices can parse the basic information standardization detection rules to obtain standardization detection rules for names, ages, and genders.

[0154] E2. The electronic device first performs a standardization check based on the standardization check rules for names. The electronic device determines the value of NotEmpty in the standardization check rules. If the value is 0, it means that there is no need to perform a standardization check on the patient's name in the electronic medical record text, and the device can directly proceed to step E3. If the value is 1, it means that a standardization check on the patient's name in the electronic medical record text is required.

[0155] When it is necessary to perform a standardization check on the patient's name in the electronic medical record text, the electronic device can search for the BText text content in the standardization check rule for the name in the medical record paragraph text corresponding to the basic information section heading of the electronic medical record text, such as "Name:". If the BText text content is not found in the medical record paragraph text, it means that the patient's name in the electronic medical record text is empty, and it can be added to the conflict prompt list, and the process jumps to step 3). If the BText text content is found in the medical record paragraph text, the device locates the end index position of the medical record paragraph text where the BText text content is found, removes the space characters, and then sequentially obtains each character after it and writes it into the variable name, until the characters contain punctuation such as {\r\n|\t|,|,|。 |;|;|} and then stops, and checks whether the variable name is empty. If name is empty, it means that the patient's name in the electronic medical record is empty, and it can be added to the conflict prompt list, and the process jumps to step E3. If name is not empty, it means that the electronic medical record text records the patient's name, and there is currently no non-standardization, so the process can continue to step E3.

[0156] E3. The electronic device can also perform standardization checks based on age-specific standardization rules. The electronic device determines the NotEmpty value in this standardization check rule. If the value is 0, it means that standardization checks on the patient's age in the electronic medical record text are not required, and the process can proceed directly to step E4. If the value is 1, it means that standardization checks on the patient's age in the electronic medical record text are required.

[0157] When it is necessary to perform a standardization check on the patient's age in the electronic medical record text, the electronic device can search for the BText text content, such as "age," in the medical record paragraph text corresponding to the paragraph title of the basic information section of the electronic medical record text, which is the standardization check rule for age. If the BText text content is not found in the medical record paragraph text, it means that the patient's age in the electronic medical record text is empty, and it can be added to the conflict prompt list, and the process jumps to step E4. If the BText text content is found in the medical record paragraph text, the device locates the end index position of the medical record paragraph text where the BText text content is found, removes the space characters, and then sequentially obtains each character after it and writes it into the variable tempAge, until the characters contain punctuation such as {\r\n|\t|,|,|。 |;|;|} and then stops, and checks whether tempAge is empty. If tempAge is empty, it means that the patient's age in the electronic medical record is empty, and it can be added to the conflict prompt list, and the process jumps to step 4). If tempAge is not empty, the electronic device can continue to search for the index position of the age unit "year", "month" or "day" in tempAge, and can be divided into the following three cases:

[0158] Case 1: All three index positions are <0. This indicates that there is no corresponding age unit in tempAge, and the patient's age in the electronic medical record text is not standardized. This can be added to the conflict warning list, and the process can proceed to step 4.

[0159] The second scenario: One index position is >0, and the other two index positions are <0. This indicates that the age unit at the index position >0 is the patient's age unit in the electronic medical record text. The electronic device can use a regular expression to obtain the value preceding this age unit in tempAge, thereby obtaining the patient's age.

[0160] The third scenario: There are two index positions > 0. The electronic device can take the smaller age unit (i.e., the earlier age unit) from these two index positions as the patient's age unit in the electronic medical record text, and use a regular expression to obtain the value preceding that age unit in tempAge, thus obtaining the patient's age.

[0161] After obtaining the patient's age, the electronic device needs to determine the age range. Here, the electronic device can process this by referring to the value range of the corresponding age unit in the standardized age detection rules. If the obtained age is not within the corresponding age unit's value range, it indicates an anomaly, which can be added to the conflict alert list, and the process can proceed to step E4. If the obtained age is within the corresponding age unit's value range, it means that the patient's age in the electronic medical record text is not non-standard, and step E4 can continue.

[0162] E4. The electronic device can also perform standardization checks based on gender-specific standardization rules. The electronic device determines the NotEmpty value in this standardization check rule. If the value is 0, it means that standardization checks on the patient's gender in the electronic medical record text are not required, and the standardization check based on the basic information standardization rules can be terminated. If the value is 1, it means that standardization checks on the patient's gender in the electronic medical record text are required.

[0163] When it is necessary to perform gender standardization checks on the patient's gender in electronic medical record text, the electronic device can search for the BText text content in the gender standardization check rule within the medical record paragraph text corresponding to the basic information section heading of the electronic medical record text, such as "Gender:". If the BText text content is not found in the medical record paragraph text, it means that the patient's gender in the electronic medical record text is empty, and it can be added to the conflict warning list, and the standardization check based on the basic information standardization check rule ends. If the BText text content is found in the medical record paragraph text, the device locates the end index position of the medical record paragraph text where the BText text content is found, removes whitespace characters, and then sequentially obtains each subsequent character and writes it to the variable sex, until the characters include punctuation such as {\r\n|\t|,|,|。 |;|;|}, and then stops, and checks whether the variable sex is empty. If sex is empty, it means that the patient's gender in the electronic medical record text is empty, and it can be added to the conflict warning list, and the standardization check based on the basic information standardization check rule ends. If sex is not empty, we can determine whether the content of sex contains the text "male" and "female", and specifically divide it into the following three cases:

[0164] The first scenario: The text contains either "male" or "female". In this case, the electronic device can determine that the patient's gender in the electronic medical record text is not irregular.

[0165] The second scenario: The text does not contain either "male" or "female". In this case, it indicates that the patient's gender is not clearly stated in the electronic medical record text. This can be added to the conflict alert list, and the standardization check based on the basic information standardization check rules can be terminated.

[0166] The third scenario: The text contains both "male" and "female". In this case, the patient's gender in the electronic medical record text cannot be determined. It can be added to the conflict alert list, and the standardization check based on the basic information standardization check rules can be terminated.

[0167] It should be noted that the above order is only an example; when electronic devices perform standardization checks based on the standardization rules for medical record paragraph text content, the order of steps E2, E3, and E4 is not required.

[0168] For normative checks based on segmented heading integrity detection rules:

[0169] The electronic device can iterate through and extract the text of each medical record segment, and search and count the segment headings corresponding to each segment to obtain the number of segment headings for that medical record segment. Specifically, it can be divided into three cases:

[0170] The first scenario: The number of section headings is greater than one. This indicates that the corresponding section heading paragraph appears repeatedly in the electronic medical record text, and can be added to the conflict alert list.

[0171] The second scenario: The number of section headings is 0. This indicates that the corresponding section headings are missing in the electronic medical record text, and can be added to the conflict alert list.

[0172] The third scenario: The number of segment headings is 1. In this case, the electronic device can further determine whether the traversed medical record segment text is empty. If it is empty, it means that the corresponding segment content in the electronic medical record text is missing, which can be added to the conflict warning list, and the standardization check based on the segment heading integrity detection rule ends. If it is not empty, it means that there are no non-standard situations, and traversal can continue.

[0173] For standardization checks based on the rules for standardization of medical record paragraph text content:

[0174] F1. The electronic device can parse the content standardization detection rules for medical record paragraphs, obtaining content standardization detection rules for different medical record paragraphs. For example, it can obtain content standardization detection rules for the "chief complaint" medical record paragraph. Simultaneously, the electronic device can also obtain a disease synonym dictionary and a list of negative words from a medical knowledge base.

[0175] F2. Electronic devices can perform standardization checks on the "chief complaint" medical record text, specifically including the following two aspects of standardization checks:

[0176] The first aspect involves the standardization check of the content length in the "chief complaint" medical record paragraph. The electronic device can first remove whitespace characters from the "chief complaint" medical record paragraph, calculate the content length, and compare it with the Length value specified in the content standardization check rules for the "chief complaint" medical record paragraph. If the content length of the "chief complaint" medical record paragraph in the electronic medical record exceeds this Length value, it can be added to the conflict alert list.

[0177] The second aspect involves text compliance checks on the "chief complaint" medical record section based on TextLimit. The electronic device can determine whether the "chief complaint" medical record section contains a disease name from a disease thesaurus; if so, it can be added to the conflict alert list. Additionally, the electronic device can determine whether the "chief complaint" medical record section contains a negative word from a negative word list; if so, it can be added to the conflict alert list. Furthermore, the electronic device can further determine whether the "chief complaint" medical record section contains content enclosed in double quotes in TextLimit (such as "several days"); if so, it can be added to the conflict alert list.

[0178] In some embodiments, the data displayed on the front-end interface by the electronic device through normative detection and conflict detection has two data types: prompt type and conflict-exclusive type. The prompt type indicates that the data is used to warn or prompt relevant information, and is generally a single data entry; the conflict-exclusive type generally corresponds to at least two data entries in the electronic medical record text, and is used to indicate that these at least two data entries are contradictory and mutually exclusive.

[0179] In some examples, the non-standard data detected by electronic devices through standardization testing are all classified as prompts.

[0180] In other examples, the collision data detected by the electronic device through collision detection includes:

[0181] The data type of the conflict data obtained from the conflict master table and conflict sub-table is determined by the configuration of the conflict master table and conflict sub-table; that is, when the electronic device performs conflict detection based on the conflict master table and conflict sub-table, it can determine whether the detected conflict data is the prompt type data or the conflict mutual exclusion type data according to the configuration of the conflict master data and the corresponding conflict sub-data.

[0182] Conflict data obtained from early warning scoring rules, surgical anesthesia risk detection rules, and drug sensitivity analysis detection rules all fall under the alert category.

[0183] Conflicting data obtained based on duplicate detection rules for positive and negative element lists all belong to the conflict-exclusive type.

[0184] It is understandable that non-standard data from different groups, as well as conflicting data from different groups, can be marked with different colors (i.e., different highlight colors are used); at the same time, different data in the same group of conflicting data can be marked with the same color (i.e., the same highlight color is used) so that users can easily see the relationship between them.

[0185] As can be seen from the above, in this embodiment, after obtaining the patient's electronic medical record text, the electronic device can perform conflict detection on the electronic medical record text based on preset conflict detection configuration information. Based on the conflict detection results, the electronic device can identify conflicting data in the electronic medical record text. This process does not rely on manual inspection of the electronic medical record by medical personnel, which can improve detection efficiency to a certain extent. Furthermore, users (e.g., medical personnel and / or medical institutions) can adjust the conflict detection configuration information according to their own habits or needs, which can further improve the accuracy of this application in detecting conflicting data in electronic medical records.

[0186] Corresponding to the medical text conflict detection method provided above, this application also provides an electronic device. This electronic device can be a server, personal computer, smartphone, or medical assistive device, or any other device with data processing capabilities; the specific type of this electronic device is not limited here. Please refer to... Figure 2 The electronic device 2 in this embodiment includes: a memory 201, and one or more processors 202. Figure 2 Only one example is shown in the diagram, along with a computer program stored in memory 201 and executable on the processor. Memory 201 stores software programs and units. The processor 202 executes various functional applications and data processing by running the software programs and units stored in memory 201 to obtain resources corresponding to preset events. Specifically, the processor 202 implements the various steps in the method embodiment by running the computer program stored in memory 201. For details, please refer to the preceding description of the method embodiment; further elaboration is omitted here.

[0187] It should be understood that, in the embodiments of this application, the processor 202 may be a central processing unit (CPU), but it may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0188] Memory 201 may include read-only memory and random access memory, and provides instructions and data to processor 202. Some or all of memory 201 may also include non-volatile random access memory. For example, memory 201 may also store device type information.

[0189] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the division of functional units and modules is only used as an example. In practical applications, functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0190] In the embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0191] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of external device software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0192] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection of apparatus or units, and may be electrical, mechanical, or other forms.

[0193] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0194] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the embodiments of this application can also be implemented by a computer program instructing associated hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of each method embodiment. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable storage medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer-readable storage devices, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable storage medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electrical carrier signals and telecommunication signals.

[0195] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for detecting conflicts in medical texts, characterized in that, include: Obtain the patient's electronic medical record text; The electronic medical record text is subjected to conflict detection based on the preset conflict detection configuration information; Based on the results of conflict detection, identify conflicting data in the electronic medical record text; The conflict detection of the electronic medical record text based on preset conflict detection configuration information includes: The patient's target basic information is obtained from the electronic medical record text, including the patient's age and gender. The conflict detection configuration information is cleaned based on the target basic information to obtain the target conflict detection configuration information; Using each medical record paragraph text as the detection object, and / or using each entity list as the detection object, conflict detection is performed based on the target conflict detection configuration information, wherein the medical record paragraph text is obtained based on the splitting operation of the electronic medical record text, and the entity list is obtained based on the extraction operation of the electronic medical record text. The preset conflict detection configuration information includes: a conflict master table and conflict sub-tables; wherein, the conflict master table is used to configure multiple conflict master data, and the conflict sub-tables are used to configure the conflict sub-data corresponding to each conflict master data; the step of cleaning the conflict detection configuration information according to the target basic information to obtain target conflict detection configuration information includes: The conflict detection configuration information is cleaned based on the target basic information and the medical entity type to which each entity list belongs, to obtain the target conflict detection configuration information; The target conflict detection configuration information includes: a target conflict master table and target conflict sub-tables; the step of using each of the medical record paragraph texts as the detection object, and / or using each of the entity lists as the detection object, based on the target conflict detection configuration information, includes: Iterate through the target conflict master table; Get the target conflict sub-table corresponding to the currently traversed conflict master data; The target conflict sub-table is matched with each of the entity lists and / or each of the medical record paragraph texts to determine whether a conflict exists; The preset conflict detection configuration information includes: early warning scoring rules, which are used to configure at least one type of key item and the scores corresponding to the numerical range of each type of key item. The early warning scoring rules include: MEWS early care early warning scoring rules for adults and PEWS early care early care early warning scoring rules for children. Correspondingly, the target conflict detection configuration information includes: target early warning scoring rules. The conflict detection based on the target conflict detection configuration information, using each of the medical record paragraph texts as the detection object and / or using each of the entity lists as the detection object, includes: The target early warning scoring rules are iterated through; Match the currently traversed key items with the text of each of the aforementioned medical record paragraphs; Based on the matching results and the scores corresponding to the numerical range of the key items, determine the patient's score on the currently visited key item; The patient's scores on each of the key items are summed to determine the results of the conflict detection.

2. The medical text conflict detection method as described in claim 1, characterized in that, The medical text conflict detection method also includes: Based on the preset standardization detection configuration information, standardization detection is performed on each segment title and each medical record paragraph text; Based on the results of the standardization test, non-standard data in the electronic medical record text is identified.

3. The medical text conflict detection method as described in any one of claims 1 to 2, characterized in that, The step of identifying conflicting data in the electronic medical record text based on the conflict detection results includes: Based on the results of the conflict detection, at least one set of conflicting data is determined in the electronic medical record text, and each set of conflicting data contains at least one piece of data; In the electronic medical record text, the at least one set of conflicting data is identified by highlighting color, wherein different sets of conflicting data use different highlighting colors, and conflicting data in the same set use the same highlighting color.

4. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 3.

5. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 3.

Images