Medical record quality control method based on medical record segmentation and scene quality control rule matching, electronic equipment and storage medium

Abstract

The invention discloses a medical record quality control method based on medical record segmentation and scenarized quality control rule matching, electronic equipment and a storage medium, the method comprises the following steps: S1, collecting patient data associated with a medical record to be subjected to quality control, and carrying out standardized preprocessing to obtain a medical record document which can be processed by a large model; s2, segmenting the medical record document according to key semantics to obtain a plurality of segmented key semantic units; s3, performing quality control rule matching on each key semantic unit to form second structured data including key semantics, unit medical record documents and quality control rules; and S4, inputting the second structured data into the large model, instructing to perform quality control, and outputting a quality control result. For medical record quality control, the accuracy of quality control rule matching is ensured through key semantic segmentation and scene quality control rule matching, quality control is effectively emphasized in combination with a large model technology, a quality control result and a processing suggestion are obtained, and powerful support is provided for subsequent medical record modification.

Description

Technical Field

[0001] The present application provides a medical record quality control method, electronic equipment and storage medium based on medical record segmentation and scenario-based quality control rule matching, relating to the field of medical information technology. Background Art

[0002] Medical records are the sum of text, symbols, charts, images, slices, and other materials generated by medical personnel during medical activities. They record the occurrence, development, diagnosis, treatment, and outcome of patients' diseases, and have both clinical summary, legal basis, and teaching and scientific research value. With the development and promotion of medical information technology, electronic medical records have gradually become popular in various medical institutions. Compared with traditional paper medical records, electronic medical records are convenient to write, easy to interact with, and flexible to access. Their emergence facilitates medical personnel's writing and query, improving work efficiency. However, they also reveal some medical record quality issues, including incomplete, erroneous, duplicated, and non-standard medical records. How to manage the quality of electronic medical records has become a new topic for hospital quality management departments.

[0003] Medical record quality control is a core component of healthcare quality management, directly impacting clinical decision-making, handling medical disputes, and the rationality of medical insurance payments. Existing quality control systems rely heavily on manual review, requiring physicians to manually check medical records for completeness, logic, and standardization, resulting in low efficiency. Hospital-level final quality control often relies on random spot checks, which cannot monitor all medical records and exhibits significant lags, making it difficult to correct errors in real time. Different departments have different understandings of quality control rules, leading to frequent homogeneity issues (e.g., gender conflicts and missing diagnostic evidence caused by the application of medical record templates). Existing systems are mostly based on simple rule bases (e.g., mandatory item checks and time limit monitoring), lacking in-depth verification of diagnosis and treatment logic and consistency of medical knowledge. For example, the system cannot identify connotation problems such as "no record of vital sign changes after surgery"; electronic medical records have the phenomenon of "putting the cart before the horse" due to the application of templates (such as "postpartum care" records for male patients), which significantly reduces the credibility of medical records; data silos, HIS, LIS, EMR and other system data are not connected, resulting in test results being disconnected from medical records and missing abnormal values; irregular signing of informed consent forms (such as signing on behalf of others, unconfirmed modifications), and inconsistencies in medical record timestamps have become major loopholes in evidence in medical disputes; the existing system only prompts the error type, and does not provide correction suggestions or related medical evidence, resulting in inefficient repeated revisions by doctors.

[0004] The current quality control system mainly relies on a rule engine based on Boolean logic. Its application scope is limited to formal verification (such as field integrity verification and timing compliance detection), and there are significant defects in quality control at the clinical semantic level. For connotation quality control scenarios involving clinical diagnosis and treatment logic (such as missing records of postoperative vital signs monitoring, drug interaction warnings, etc.), traditional systems lack the reasoning ability based on evidence-based medicine. Especially for practicing physicians with little clinical experience, existing technologies are unable to build a closed-loop intelligent assistance system, and it is difficult to provide clinically operational improvement suggestions in dimensions such as document standardization guidance, differential diagnosis prompts, and treatment plan rationality evaluation. Summary of the Invention

[0005] The technical problem to be solved by this application is that the variety of medical record types makes it difficult to accurately perform medical record quality control.

[0006] To solve the above technical problems, the technical solution of the present application provides a medical record quality control method based on medical record segmentation and scenario-based quality control rule matching, including the following steps:

[0007] S1 collects patient data associated with the medical records to be quality-controlled, performs standardized preprocessing, and obtains medical records that can be processed by the large model;

[0008] S2: segmenting the medical record document according to key semantics to obtain a plurality of segmented key semantic units, wherein the key semantic units are set as first structured data including the key semantics and the unit medical record document;

[0009] S3 matches quality control rules for each key semantic unit to form the second structured data including key semantics, unit medical record documents, and quality control rules;

[0010] S4 inputs the second structured data into the large model, instructs quality control, and outputs the quality control results.

[0011] Preferably, the preprocessing in step S1 includes preprocessing operations of de-anonymizing, cleaning and word segmentation of patient data; including using a large language model to perform semantic analysis on the medical record text in the patient data, extract key diagnosis and operation information, and form structured data.

[0012] Preferably, the pre-processing in step S1 further includes adding corresponding content type numbers to different text contents in the medical record document to form secondary processed structured data.

[0013] Preferably, the key semantic types include chief complaint, current medical history, auxiliary examination conclusion, differential diagnosis basis, diagnosis, and treatment plan; the first structured data structure form is set as:

[0014]

Key semantics

[0015]

Medical record paragraph

[0016] Preferably, step S3 includes:

[0017] S31 first matches the scenario-based quality control rule library for key semantic units;

[0018] S32 then matches the quality control rules in the scenario-based quality control rule library;

[0019] S33 incorporates the matched quality control rules into the first structured data to obtain second structured data:

[0020]

Key semantics

[0021]

Medical Record Document Paragraph

[0022]

Quality control rules

[0023] Preferably, step S4 includes:

[0024] Generate instructions for inputting into the large model based on the second structured data and the instruction template, instruct the large model to perform quality control and output quality control results, wherein the quality control results include qualified, unqualified, and unqualified processing suggestions.

[0025] Preferably, the instruction template is set to:

[0026]

[0027] Preferably, the large model is set as a fine-tuned large model based on the medical record quality control data set.

[0028] The present application also provides an electronic device, comprising:

[0029] A memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program to implement the aforementioned medical record quality control method based on medical record segmentation and scenario-based quality control rule matching.

[0030] The present application also provides a computer-readable storage medium for storing a computer program, which, when executed, implements the aforementioned medical record quality control method based on medical record segmentation and scenario-based quality control rule matching.

[0031] This application focuses on medical record quality control. It ensures the accuracy of quality control rule matching through key semantic segmentation and scenario-based quality control rule matching. Combined with large model technology, it effectively focuses on quality control, obtains quality control results and processing suggestions, and provides strong support for subsequent medical record modifications. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A flow chart of the medical record quality control method provided in an embodiment of the present application;

[0033] Figure 2 This is a diagram for data structuring processing of medical record documents;

[0034] Figure 3 A schematic diagram of the specific steps for matching quality control rules;

[0035] Figure 4 Schematic diagram of fine-tuning a large model using the LoRA method. DETAILED DESCRIPTION

[0036] To make the present application more obvious and easy to understand, various exemplary embodiments will be introduced below. These examples are non-limiting, and it should be understood that they are used to illustrate the broader application aspects of the device, system and method. Without departing from the essence and scope of the present application, these embodiments can be subjected to various changes and can be replaced by equivalents. In addition, various changes can be carried out to adapt to the purpose, content or scope of the present application in order to adapt to special circumstances, materials, material components, treatment types, treatment actions or steps. All of these changes will be within the scope of protection of the present application.

[0037] Any materials, dimensions, or quantities described in the overview or detailed description are intended to be examples only and are not intended to limit the subject matter of this application. Furthermore, the various implementations of the embodiments described herein are intended to complement each other, rather than to replace each other, unless otherwise indicated. In other words, implementations from one embodiment can be freely combined with implementations from other embodiments, as those skilled in the art will readily appreciate, unless otherwise indicated.

[0038] In the description of this application, it should be noted that the terms "inner" and "outer" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, or the orientations or positional relationships in which the product of this application is typically placed when in use. These terms are intended solely to facilitate the description of this application and simplify the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first" and "second" and the like are used solely for distinction and should not be construed as indicating or implying relative importance.

[0039] It should also be noted that, in the description of this application, unless otherwise expressly specified or limited, the terms "disposed" and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to direct connections, indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in this application based on the specific circumstances.

[0040] Example 1

[0041] The present application embodiment provides a medical record quality control method based on medical record segmentation and scenario-based quality control rule matching, see Figure 1 , specifically including the following steps:

[0042] S1 collects patient data associated with the medical records to be quality controlled, performs standardized preprocessing, and obtains medical record documents that can be processed by the large model; since the patients associated with the medical records to be quality controlled have a lot of in-hospital data, such as medical records, test reports, examination reports, medical order records, etc., preprocessing is required to obtain standardized data.

[0043] Exemplarily, the preprocessing operations include de-identification, cleaning, and word segmentation of patient data; they may also include using a large language model to perform semantic analysis on the medical record text in the patient data, extract key diagnosis and operation information, and form structured data.

[0044] In a further embodiment, after step S1 performs standardized preprocessing on the patient data, a medical record document that can be processed by the large model in the form of text content is obtained, and in the process of processing the medical record document, corresponding content type numbers are added according to different text contents such as test, examination report, medication data, etc. Figure 2 , forming structured data for secondary processing, which facilitates the retrieval, query, extraction and processing of medical records in subsequent steps.

[0045] The medical data ecosystem has significant heterogeneity, encompassing multi-source heterogeneous data such as electronic health records (EHRs), laboratory test reports, medical imaging diagnoses, and drug treatment plans. These data present multidimensional complexity in terms of the degree of structure (structured coded data, semi-structured text, and unstructured free text), semantic expression paradigm (medical terminology, clinical narrative language), and data granularity (from molecular-level test indicators to organ system-level descriptions). Achieving precise quality control requires the establishment of a multimodal data fusion framework, involving key technologies such as natural language processing (NLP)-driven entity recognition and relationship extraction, mapping to international standardized medical terminology systems (such as SNOMED CT and ICD), and semantic normalization processing based on knowledge graphs.

[0046] S2 divides the medical record document according to key semantics to obtain a plurality of key semantic units after segmentation, wherein the key semantic units are first structured data including key semantics and medical record document paragraphs.

[0047] Medical records to be quality-controlled generally include multiple key semantics such as chief complaint, current medical history, auxiliary examination conclusions, differential diagnosis basis, diagnosis, treatment plan, and document content corresponding to the key semantics. The quality control rules for document content corresponding to different key semantics often differ.

[0048] For example, for the chief complaint, the quality control rules require that the medical record documents accurately record the patient's chief complaint, including the onset time, symptom changes, past medical history, etc.; for the treatment plan, the quality control rules require that the medical record documents accurately record specific measures such as drug treatment, physical therapy, surgical treatment, as well as patient precautions and follow-up plans. It can be seen that there are often differences in the quality control rules for the content of documents corresponding to different key semantics. Therefore, the medical record quality control method provided in the embodiment of the present application divides the medical record documents by key semantics in step S2 to obtain multiple key semantic units after segmentation. It is understandable that the classification standards of key semantics usually need to rely on the standards of the medical community and belong to the existing technology category such as industry standards.

[0049] At the beginning of medical record quality control, a mechanism for segmenting medical record content and adapting to quality control scenarios is established. By integrating the patient's data in the hospital, the main paragraphs in the medical record content, such as chief complaint, current medical history, auxiliary examinations, preliminary diagnosis, etc., are effectively segmented. Large model technology can also be used to achieve rapid segmentation. In order to adapt to the multiple medical record templates of each hospital, an adjustable paragraph segmentation mechanism is used to adapt to the medical record templates of each hospital, ensuring the best quality control effect in the medical record quality control of each hospital.

[0050] The first structured data obtained in step S2 can be understood as data stored in a structured form in which key semantics are bound to medical record paragraphs. The content in the medical record paragraphs is content belonging to the key semantics. The first structured data:

[0051]

Key semantics

[0052]

Medical record paragraph

[0053] After the medical record document to be quality controlled is processed in step S2, the content of the medical record document is stored in the first structured data format, which can be easily and quickly queried and retrieved, and is convenient for subsequent operations.

[0054] S3 matches quality control rules for each key semantic unit to form second structured data including key semantics, unit medical record documents, and quality control rules.

[0055] In a further embodiment, see Figure 3 , step S3 specifically includes:

[0056] S31 first matches the scenario-based quality control rule base for the key semantic unit. The reason for establishing a scenario-based quality control rule base is that according to existing specifications, there are differences in quality control rules in different scenarios, but the same key semantics may appear in different scenarios. Therefore, for the key semantic unit, in addition to matching the quality control rules corresponding to the key semantics, it is also necessary to consider the differences in quality control rules under the scenario. Therefore, the embodiment of the present application adopts the method of establishing a scenario-based quality control rule base. When matching quality control rules for each key semantic unit, the scenario-based quality control rule base is matched first, and then the matching quality control rules are queried from the matching scenario-based quality control rule base.

[0057] S32 then matches the quality control rules in the scenario-based quality control rule library;

[0058] S33 incorporates the matched quality control rules into the first structured data to obtain second structured data:

[0059]

Key semantics

[0060]

Medical Record Document Paragraph

[0061]

Quality Control Rules

[0062] There may be one or more matching quality control rules.

[0063] For example, the scenario-based quality control rule library has different quality control rules according to the medical records in different scenarios. For example, the inpatient medical records, outpatient medical records and special disease medical records generated in the inpatient scenario, outpatient scenario and special disease scenario have different corresponding scenario-based quality control rule libraries; therefore, it is necessary to conduct advance collection and preparation work and build a scenario-based quality control rule library based on existing medical quality control standards.

[0064] Hospitalization scenario - scenario-based quality control rule library (not exhaustive) includes:

[0065] Hospitalization scenario-scenario-based quality control rule library Chief complaint-related quality control rules Current medical history-related quality control rules Personal History-Related Quality Control Rules Family History-Related Quality Control Rules Physical Examination-Related Quality Control Rules Medical records-related quality control rules Ward rounds records-related quality control rules ... .

[0066] Outpatient Scenarios - Scenario-based quality control rule base (not exhaustive) includes:

[0067] Outpatient Scenario-Scenario-Based Quality Control Rule Library Patient basic information-related quality control rules Chief complaint-related quality control rules Current medical history-related quality control rules Physical Examination-Related Quality Control Rules Auxiliary inspection-related quality control rules Diagnosis-related quality control rules Treatment plan-related quality control rules ... .

[0068] Special disease scenarios - scenario-based quality control rule library (not exhaustive list) includes:

[0069]

[0070] The specific content of quality control rules comes from medical standards and falls within the existing technical scope such as industry standards.

[0071] S4 inputs the second structured data into the large model, instructs quality control, and outputs the quality control results.

[0072] Step S3 obtains a plurality of second structured data, and generates instructions for inputting into the large model according to the second structured data and the instruction template, instructing the large model to perform quality control and output quality control results, wherein the quality control results include qualified, unqualified and unqualified processing suggestions.

[0073] Exemplarily, the instruction template applied to the second structured data is:

[0074]

[0075] Batch quality control can be performed on the batch second structured data obtained in step S3. The quality control results can be summarized to generate a result determination table for this quality control medical record, in which whether each item is qualified or unqualified is recorded; for unqualified items, the processing suggestions can be summarized to obtain a result review and processing suggestion table; the above result determination table and result review and processing suggestion table constitute the medical record quality control output results for medical staff or management personnel to review. The processing suggestions include modification suggestions, which provide strong support for subsequent medical record modifications.

[0076] Limited by the inherent characteristics of the Transformer architecture, existing large models face a dual challenge when processing extremely long clinical texts (such as critical care records containing the complete course of a disease): First, the limited model window length leads to the loss of time series information of key medical events; second, the computational complexity of the self-attention mechanism is proportional to the square of the sequence length, making the model prone to semantic focus drift when processing long texts. This technical limitation directly leads to systematic errors in medical record quality assessment, such as the omission of key indicators and the failure of outlier detection, seriously affecting the clinical credibility of quality control results.

[0077] In the embodiment of the present application, a set of multi-rule adaptation methods based on multiple medical record types is designed for the multiple rules and types in common quality control medical records. By grouping and segmenting the contents of the current quality control medical record, matching the quality control rules of the current medical record type, adapting the quality control processing logic of the rule and the module content contained therein, batch rule quality control is performed after obtaining the rules. Combined with large model technology and the module content in the rules, effective focus on quality control is performed to obtain quality control results and modification suggestions, original content and location of the medical record and other key information, providing strong support for subsequent medical record modifications.

[0078] It can be understood that the medical record quality control method based on medical record segmentation and scenario-based quality control rule matching provided in the embodiment of the present application, wherein the implementation of all or part of the process relies on computer programs to deploy or instruct corresponding hardware such as storage devices and processors to complete. After reading and understanding all or part of the process of the medical record quality control method based on medical record segmentation and scenario-based quality control rule matching provided in the embodiment of the present application, ordinary technicians in this field can easily implement it through computer programs. There are no technical barriers for ordinary technicians in this field and no creative labor is required.

[0079] In a further embodiment, the large model for quality control is a fine-tuned large model. Specifically, the large model is fine-tuned through a medical record quality control data set. The medical record quality control data set includes collected medical record documents, quality control rules, and manual quality control results. Medical experts perform manual quality control to ensure the accuracy of the data set.

[0080] The fine-tuning method can use the LoRA method, based on the open source large language model Alibaba Cloud's open source Tongyi Qianwen Qwen-14B, see Figure 4 , the LoRA (low rank adaptation) method is used for efficient fine-tuning. The LoRA method refers to the use of low-rank decomposition technology to achieve efficient parameter updates without modifying the original model parameters. The specific implementation methods include: model architecture design, adding a dimensionality reduction-dimensionality increase structure in parallel to each layer of the open source large language model bypass. The structure consists of two serial linear transformation layers, the dimensionality reduction layer (matrix A), projects the input from the high-dimensional space (dimension d, usually ≥1024) to the low-dimensional space (dimension r, r<<d); the dimensionality increase layer (matrix B), maps the features from the low-dimensional space r back to the original dimension d; parameter initialization strategy, matrix A is initialized with a random Gaussian distribution, and matrix B is initialized to a zero matrix to ensure that the bypass output is zero in the initial stage of training; computational advantage, traditional fine-tuning requires updating the d×d parameter matrix, and LoRA decomposes it into the product of A(d×r) and B(r×d), and the total number of parameters is reduced from d 2 When r < < d, the computational resource requirements are significantly reduced. This method introduces a low-rank adapter, which allows effective model fine-tuning by training only a small number of new parameters while keeping the original model parameters frozen. This approach is particularly suitable for resource-constrained applications. The medical record quality control dataset is an annotated set obtained through rigorous manual verification. Fine-tuning the large model based on this dataset, used for medical record quality control, produces more accurate quality control results than general-purpose large models.

[0081] The fine-tuning method can also adopt a combination of incremental learning + reinforcement fine-tuning and supervised fine-tuning (IL+RLFT, SFT) to fine-tune the large model through the medical record quality control dataset; through incremental learning (IL), on the basis of the open source model, the model is incrementally learned using the labeled medical record quality control dataset to adjust the model parameters to adapt to the new data distribution, and the new dataset provided by the medical record quality control dataset is merged into the existing model, which can improve the quality control ability of the large model in medical record quality control. The quality control results of the fine-tuned large model are more reasonable and more in line with medical scenarios.

[0082] Example 2

[0083] In another embodiment, steps S3 and S4 are specifically as follows, based on the consideration of the scenario preconditions for measuring the correspondence between the scenario and the rule:

[0084] S3 medical records match quality control rules, including:

[0085] S31 matches the type of the current quality control medical record document with the scenario-based rule library. For example, the admission record only matches the scenario-based rule of the admission record type. The matched scenario-based rule is combined with the current medical record document to assemble into text content that can be processed by the large model. The text is input to the large model, instructing the large model to perform rule matching and outputting a condition and result table corresponding to the medical record document and the scenario-based rule. The scenario-based rule is updated to the structured data corresponding to the current medical record document based on the condition and result table. The condition refers to the condition for executing the scenario-based rule, also known as the scenario execution condition or scenario precondition.

[0086] S32 performs multi-rule matching on the medical records that currently require quality control, and matches the quality control rules that match the current medical record document type; as a supplement to step S31, step S32 performs multi-scenario and multi-rule matching to ensure that comprehensive quality control rules are matched.

[0087] S4 inputs medical records into the big model, instructs quality control, and outputs quality control results; specifically, it includes:

[0088] Before executing the matched scenario-based rules, S41 retrieves the scenario-based results from step S31 and filters out the quality control rules that need to be executed based on the scenario-based rule execution conditions. For example, if the scenario execution conditions are not met, the quality control of the current rule will not be performed. If the scenario execution conditions are met, the quality control of the current rule will be performed. Therefore, there is a one-to-one correspondence between the scenario execution conditions and the rules, and the scenario execution conditions can be inferred from the rules.

[0089] S42 continues to extract the logical operation statement corresponding to the rule content after the above filtering, and performs corresponding variable replacement. For example, if the current quality control rule mentions the patient's examination report data, the patient's already processed examination report data needs to be replaced. The same is true for other types of data.

[0090] S43 inputs the replaced logical operating language into the large model, and outputs the result judgment table of this quality control medical record, as well as the result review and processing suggestion table for the results judged as unqualified.

[0091] Based on this embodiment, this application provides a multi-scenario mode adapter. For some complex quality control situations, a set of adaptation mechanisms for complex quality control scenarios is designed. By pre-fabricating pre-scenarios and combining the text processing capabilities of large models, it is possible to judge and analyze whether the quality control medical record content meets the quality control scenario preconditions in the quality control scenario.

[0092] Example 3

[0093] In another embodiment, the present application further provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and when the processor executes the computer instructions, implements the medical record quality control method based on medical record segmentation and scenario-based quality control rule matching provided in this application.

[0094] An embodiment of the present application also provides an electronic device, specifically including: a processor and a memory, wherein the memory is used to store a computer program, and the computer program is loaded and executed by the processor to implement a medical record quality control method based on medical record segmentation and scenario-based quality control rule matching.

[0095] The above description is only a preferred embodiment of the present application and does not limit the present application in any form or substance. It should be pointed out that ordinary technicians in this technical field can make several improvements and supplements without departing from the present application, and these improvements and supplements should also be regarded as the scope of protection of the present application. Any technician familiar with this profession can make some changes, modifications and equivalent changes made by using the technical content disclosed above without departing from the content and scope of the present application, which are all equivalent embodiments of the present application; at the same time, any equivalent changes, modifications and evolutions made to the above-mentioned embodiments based on the essential technology of the present application are still within the scope of the technical solution of the present application.

Claims

1. A medical record quality control method based on medical record segmentation and scenario-based quality control rule matching, characterized by: The following steps are involved: S1 collects patient data associated with the medical records to be quality-controlled, performs standardized preprocessing, and obtains medical records that can be processed by the large model; S2: segmenting the medical record document according to key semantics to obtain a plurality of segmented key semantic units, wherein the key semantic units are set as first structured data including the key semantics and the unit medical record document; S3 matches quality control rules for each key semantic unit to form the second structured data including key semantics, unit medical record documents, and quality control rules; S4 inputs the second structured data into the large model, instructs quality control, and outputs the quality control results.

2. A medical record quality control method based on medical record segmentation and scenario-based quality control rule matching according to claim 1, characterized in that: The preprocessing in step S1 includes the preprocessing operations of de-privacy, cleaning and word segmentation of patient data; including the use of a large language model to perform semantic analysis on the medical record text in the patient data, extract key diagnosis and operation information, and form structured data.

3. A medical record quality control method based on medical record segmentation and scenario-based quality control rule matching according to claim 2, characterized in that: The pre-processing in step S1 also includes adding corresponding content type numbers to different text contents in the medical record document to form secondary processed structured data.

4. The medical record quality control method based on medical record segmentation and scenario-based quality control rule matching according to claim 1 is characterized in that: The key semantic types include chief complaint, current medical history, auxiliary examination conclusion, differential diagnosis basis, diagnosis, and treatment plan; the first structured data structure form is set as: 【Key semantics】 【Medical record paragraph】.

5. The medical record quality control method based on medical record segmentation and scenario-based quality control rule matching according to claim 1 is characterized in that: Step S3 includes: S31 first matches the scenario-based quality control rule library for key semantic units; S32 then matches the quality control rules in the scenario-based quality control rule library; S33 incorporates the matched quality control rules into the first structured data to obtain second structured data: 【Key semantics】 【Medical Record Document Paragraph】 【Quality control rules】 6. The medical record quality control method based on medical record segmentation and scenario-based quality control rule matching according to claim 1 is characterized in that: Step S4 includes: Generate instructions for inputting into the large model based on the second structured data and the instruction template, instruct the large model to perform quality control and output quality control results, wherein the quality control results include qualified, unqualified, and unqualified processing suggestions.

7. The medical record quality control method based on medical record segmentation and scenario-based quality control rule matching according to claim 6 is characterized in that: The instruction template is set as:

8. The medical record quality control method based on medical record segmentation and scenario-based quality control rule matching according to claim 6 is characterized in that: The large model is set as a fine-tuned large model based on the medical record quality control dataset.

9. An electronic device, characterized in that: include: A memory and a processor; the memory is used to store a computer program; The processor is used to execute the computer program to implement the medical record quality control method based on medical record segmentation and scenario-based quality control rule matching as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that Used to store a computer program, which, when executed, implements the medical record quality control method based on medical record segmentation and scenario-based quality control rule matching as described in any one of claim 8.

Images