Clinical medicine graduation practice teaching management system

Abstract

The invention, which relates to the technical field of medical teaching management, discloses a clinical medical graduation practice teaching management system comprising a memory, a theoretical data integration module, a clinical data integration module and an intelligent evaluation module. Through combination of theoretical test paper and theoretical operation tests, a comprehensive score is calculated, the mastering degree of the intern on knowledge points of all departments is effectively measured, knowledge units needing to be strengthened are automatically recognized according to a set qualified threshold value, a training plan is dynamically adjusted, and it is ensured that the intern enters the next rotation department after mastering all key knowledge points; the attendance condition is automatically identified by combining attendance scheduling and actual attendance records, and shift supplementation is arranged; respectively analyzing the logicality and accuracy according to the electronic medical record and the clinical practice video to obtain a logicality score and an accuracy score; clinical data of interns are integrated from multi-dimensional data such as electronic medical records, attendance records and clinical practice videos, and visual learning progress analysis is provided for teaching teachers.

Description

Technical Field

[0001] This invention relates to the field of medical teaching management technology, specifically a clinical medical graduation internship teaching management system. Background Technology

[0002] Clinical medicine graduation internship teaching management refers to the management of clinical medicine interns by supervising teachers during their graduation internship. This usually includes internship planning, supervision of the teaching process, assessment and evaluation, and medical and patient safety protection, to ensure that interns accurately master clinical skills, cultivate professional qualities, and smoothly transition from theoretical learning to clinical practice.

[0003] The current clinical medical graduation internship teaching management system is mainly limited to assessing the attendance of interns. It usually relies on the supervising teachers to conduct practical assessments and evaluations of interns. This teaching management method lacks data statistical analysis and intelligent evaluation, making it difficult to comprehensively and scientifically grasp the specific internship situation of interns. Summary of the Invention

[0004] The purpose of this invention is to solve the problems mentioned in the background section by proposing a clinical medical graduation internship teaching management system.

[0005] The objective of this invention can be achieved through the following technical solution: a clinical medical graduation internship teaching management system, comprising: a storage device, a theoretical data integration module, a clinical data integration module, and an intelligent evaluation module;

[0006] The memory is used to integrate and store the theoretical and clinical data of interns in various rotation departments. The specific theoretical data includes the scores of theoretical test papers and theoretical video operation tests, while the clinical data includes the electronic medical records written by interns, the attendance schedule and actual attendance records, and clinical practice videos.

[0007] The theoretical data integration module integrates theoretical data from interns' rotations in various departments. Where n is a positive integer, m is the total number of rotating departments; Q is the score of the theoretical test paper, S is the score of the theoretical video operation test, and C is the number of tests;

[0008] The clinical data integration module integrates clinical data from interns' rotations in various departments. Where f is a positive integer, representing the total number of attendances, m is the total number of rotating departments; R is the logical reasoning score, and H is the accuracy score.

[0009] The intelligent assessment module performs an intelligent assessment of the intern's performance throughout the entire period based on integrated theoretical and clinical data. Specifically:

[0010] The integrated theoretical data is weighted and summed to obtain the theoretical score L for each intern in each rotation department. Therefore, the theoretical scores for each intern in each rotation department are denoted as (L1, L2, L3...L...). m );

[0011] The integrated clinical data is weighted and summed to obtain the internship scores for each intern in each rotation department. Therefore, the internship scores for each intern in each rotation department are denoted as (T1, T2, T3...T...). m );

[0012] The internship score for each intern is calculated by weighting their theoretical score (L) and practical score (T) in each rotation department, and then visualized in tabular form.

[0013] As a preferred embodiment of the present invention, the specific method for integrating theoretical data and clinical data is as follows:

[0014] The theoretical test papers and theoretical practical test videos of interns in each rotation department were collected. The supervising teachers scored the theoretical test papers and videos separately to give scores. This yielded the theoretical test paper score and theoretical practical test video score for each test in each rotation department, which were denoted as Q and S respectively. Therefore, the theoretical data for each rotation department can be calculated as {(Q1,S1),(Q2,S2),(Q3,S3)...(Q...S1)}. n ,S n )}, where Q k and S k The score is the score of any one theoretical test and theoretical practical test in the rotating department;

[0015] It connects with the hospital's HIS / EMR system to collect electronic medical records written by interns in various rotation departments, their scheduled attendance records and actual attendance records, and clinical practice videos, and records them as clinical data.

[0016] As a preferred embodiment of the present invention, the specific method for interns to integrate theoretical data from their rotations in various departments is as follows:

[0017] Step 1: The theoretical data of the interns in their rotation departments = {(Q1,S1),(Q2,S2),(Q3,S3)...(Q n ,S n )} Calculate the theoretical score F for any test (including theoretical written tests and theoretical practical tests). k The specific calculation formula is as follows:

[0018]

[0019] Where α Qα is the weighting factor for the theoretical test paper. S This is the weighting factor for the theoretical practical test; from this, the theoretical score F for each test unit in the intern's rotation department can be obtained. k ;

[0020] Step 2: Assume there is a passing score, denoted as F. 合格 And compare it with the theoretical score F of each test unit. k Compare; if the theoretical score F of the test unit k <F 合格 When this happens, the test unit is recorded as a reinforcement unit;

[0021] Step 3: For the enhanced units, implement intensive training for interns. After the intensive training is completed, conduct the theoretical paper test and theoretical video test for the enhanced units again until the theoretical score of the enhanced units is greater than or equal to the set passing score.

[0022] Step Four: When an intern passes all test units in their rotation department, they proceed to the next rotation department until all rotations are completed. If an intern passes all test units in a rotation department on their first attempt, the written test score and video operation score are used as the final score for that rotation. If an intern fails all test units in a rotation department on their first attempt, the written test score and video operation score of the last passed test are used as the final score for that rotation, and the number of tests for that unit is counted. The intern's theoretical data for that rotation department is integrated into: {(Q1,S1,C1),(Q2,S2,C2),(Q3,S3,C3)...(Q n ,S n C n )}, where C represents the number of tests; thus, the theoretical data for each rotation department throughout the entire internship period can be integrated into Where m is a positive integer, and m represents the total number of rotating departments.

[0023] As a preferred embodiment of the present invention, the specific method of reinforcement training is as follows:

[0024] Strengthen theoretical knowledge training: Use NLP to extract core knowledge points from medical literature and medical record databases, use graph databases to establish relationships between knowledge points, use similarity algorithms to match the most relevant knowledge points based on the interns' theoretical test scores, automatically extract relevant questions and answer explanations from the relevant knowledge points, dynamically construct a theoretical test to strengthen theoretical knowledge, and send it to the interns.

[0025] Strengthen theoretical and practical training: Set up a standard operation video corresponding to the theoretical operation test of the test unit, use computer vision to identify key steps, and timestamp the key steps. Then send the annotated standard operation video to the interns.

[0026] As a preferred embodiment of the present invention, the specific method of clinical data integration is as follows:

[0027] Step 1: Extract the intern's scheduled attendance record and actual attendance record. If the actual attendance record covers the scheduled attendance record, send the electronic medical record to Step 3 and the clinical practice video to Step 4; otherwise, proceed to Step 2.

[0028] Step 2: When the actual attendance records cannot cover the attendance schedule records, the uncovered part of the attendance schedule records is extracted as the make-up work hours. The supervising supervisor arranges the make-up work time according to their own time and the make-up work hours. When the intern completes the make-up work hours, return to Step 1.

[0029] Step 3: Extract the intern's electronic medical record and use NLP to perform logical analysis to obtain three elements: temporal rationality, causal coherence, and contradiction elimination rate. These three elements are denoted as (D... 时序合理度 D 因果连贯度 D 矛盾排除率 The logical reasoning score R is calculated based on the three elements. Therefore, the logical reasoning score for each electronic medical record written by the intern is denoted as (R1, R2, R3...R...). f ), where f represents the total number of attendances;

[0030] Step Four: Extract clinical practice videos and analyze the accuracy of the interns' clinical practice to obtain the two elements of practice. The two elements are: temporal alignment and gesture accuracy, denoted as (A...). 时序对齐度 A 手势精准度 The accuracy score H is calculated based on the two elements, and the accuracy score of each clinical practice video written by the intern is recorded as (H1, H2, H3...H). f );

[0031] Step 5: Integrate the clinical data from each attendance session into {(R1,H1),(R2,H2),(R3,H3)...(R f H f The clinical data of interns throughout their rotations in various departments will be integrated into a unified whole.

[0032] As a preferred embodiment of the present invention, the logical evaluation method for electronic medical records is as follows:

[0033] Step 3-1: The calculation process for timing rationality is as follows:

[0034] Collect a large number of real medical records and extract standard time sequence paths, or have the supervising teacher formulate and publish standard time sequence paths; construct a standard time sequence based on the standard time sequence paths, and record the standard time sequence as a standard process sequence B(b1,b2,b3...bn); identify event timestamps in electronic medical records, use dynamic time warping to calculate the time sequence deviation in electronic medical records, and record it as the event flow P(p1,p2,p3...pn) of electronic medical records;

[0035] The formula for calculating DTW distance and maximum possible DTW distance, and the formula for calculating time series reasonableness are as follows:

[0036] Step 3-2: The process of calculating causal coherence is as follows:

[0037] Dependency parsing was used to analyze causal sentences in electronic medical records and extract causal relationships, including the logical chain of etiology → diagnosis → treatment → prognosis. Based on the NLP processing results, the causal relationships in the medical records were organized into a sequence denoted as {(G1,E1),(G2,E2),(G3,E3)...(G... n E n In the equation, G is the cause and E is the result, and G and E constitute a causal relationship.

[0038] Using medical knowledge maps, standard causal logical relationships are obtained from medical guidelines or textbooks, forming a standard causal sequence denoted as {(g1,e1),(g2,e2),(g3,e3)...(g...}. n ,e n )}, where g is the standard cause and e is the standard outcome, and g and e constitute a standard causal relationship;

[0039] The similarity between medical record causal pairs and standard causal pairs is calculated using text similarity: Similarity{(G,E),(g,e)}. A matching threshold of 0.8 is set, and a match is considered successful when Similarity{(G,E),(g,e)} > 0.8. The number of successfully matched causal pairs is counted, and the number of standard causal relationships is calculated: Number of standard causal relationships = Total number of standard causal pairs related to the disease in the medical knowledge base.

[0040] Calculate causal coherence:

[0041] Step 3-3: The calculation process for the conflict elimination rate is as follows:

[0042] NLP is used to identify contradictions within medical records. The number of conflict detections in the medical record text is recorded as the actual number of conflicts. The total number of possible conflicts is calculated as the number of all possible conflict pairs that can be detected in the medical record.

[0043] Calculate the contradiction elimination rate:

[0044] As a preferred embodiment of the present invention, the DTW distance and the maximum possible DTW distance are calculated as follows:

[0045] Where d(s) i b j ) is the distance between two events, and W is the alignment path, i.e., how to match two sequences. The optimal alignment path is obtained through dynamic path planning calculation. The specific path planning calculation formula is:

[0046]

[0047] The maximum possible DTW distance is the maximum DTW cost when two sequences are completely misaligned or unrelated, defined by the following formula: Where i and j are any events in the standard process sequence B (b1,b2,b3...bn) and the event flow P (p1,p2,p3...pn) of the electronic medical record, respectively; This represents the event matching error.

[0048] As a preferred embodiment of the present invention, the accuracy assessment method for clinical practice videos is as follows:

[0049] Step 4-1: Analyze the timing alignment of key steps to obtain the timing alignment degree. The specific process is as follows:

[0050] Given a standard operation video, computer vision and NLP are used to analyze voice commands in the intern's clinical practice video to extract the actual operation timestamps of key steps. Dynamic time warping is employed to calculate the temporal alignment between the intern's operation and the standard operation, denoted as A. 时序对齐度 ;

[0051] Step 4-2: Analyze the gestures in key steps to obtain gesture accuracy. The specific process is as follows:

[0052] Hand key points were extracted using MediaPipe Hands / OpenPose. The key coordinates at frame i of the intern's clinical practice video and the key point coordinates at frame i of the standard operation video were compared using Euclidean distance to measure the gesture deviation. The gesture deviations of each key step were then summed to obtain the total gesture deviation, denoted as V. gesture ;

[0053] The time matching error between clinical practice videos and standard operating procedure videos is calculated using dynamic time warping and denoted as V. DTW ;

[0054] Total gesture deviation V gesture Matching error V DTW The gesture accuracy A is obtained through formulaic calculation and analysis. 手势精准度 The specific calculation formula is as follows:

[0055]

[0056] Where max(gestuer) is the maximum allowed gesture error, max(DTW) is the maximum allowed time matching error, and β gesture β is the weighting factor for the total gesture deviation. DTW The weighting factor for the time matching error is set.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] 1. This invention effectively measures an intern's mastery of knowledge points in various departments by combining theoretical exams and practical tests to calculate a comprehensive score; it uses NLP technology to extract core knowledge points, analyzes knowledge relationships using knowledge graphs, and dynamically constructs personalized exams based on test results to achieve precise reinforcement training; it automatically identifies knowledge units that need reinforcement based on set passing thresholds and dynamically adjusts the training plan to ensure that interns enter the next rotation department after mastering all key knowledge points; it records interns' test scores, number of passes, and other data, comprehensively integrating data from all stages of rotation to provide mentors with visualized learning progress analysis.

[0059] 2. This invention automatically identifies absences and arranges make-up shifts by combining scheduled attendance with actual attendance records, ensuring the integrity of internship time. It comprehensively evaluates the logicality of electronic medical records based on their temporal consistency, causal coherence, and contradiction elimination rate, obtaining a logicality score. Simultaneously, it comprehensively assesses the accuracy of interns' clinical practice by evaluating the temporal alignment and gesture precision of clinical practice videos and standard operating procedure videos, obtaining an accuracy score. This invention integrates interns' clinical data from multiple dimensions, including electronic medical records, attendance records, and clinical practice videos, providing instructors with visualized learning progress analysis.

[0060] 3. This invention comprehensively evaluates the interns' overall abilities by combining their theoretical and clinical scores in each rotation department, reducing subjective evaluation errors, obtaining the interns' internship scores in each rotation department, and using data visualization tools to present the interns' learning progress at each stage. This helps supervising teachers to accurately formulate training programs, provides data support for hospitals to optimize internship management, and improves the overall teaching quality.

[0061] In summary, this invention achieves deep integration of clinical data and teaching management through data penetration integration and intelligent dynamic evaluation, transforming the medical education model from experience-driven to data-driven. Attached Figure Description

[0062] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0063] Figure 1 This is a schematic diagram of the principle of the present invention. Detailed Implementation

[0064] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] It should be understood that the terms “comprising” and “including” used in this disclosure and claims indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0066] It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this disclosure. As used in this disclosure and claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this disclosure and claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations.

[0067] Please see Figure 1 As shown, a clinical medical graduation internship teaching management system includes: a memory, a theoretical data integration module, a clinical data integration module, and an intelligent assessment module;

[0068] Data integration was performed on interns to obtain internship data, which was then stored in memory. The specific data integration method was as follows:

[0069] The theoretical test papers and theoretical operation test videos of interns in each rotation department were collected. The theoretical operation videos refer to videos of interns practicing clinical surgery in the practice room (specifically, using VR / AR equipment to simulate clinical operations and recording this process to form theoretical operation test videos). The supervising teachers scored the theoretical test papers and theoretical operation videos separately to give scores. Thus, the theoretical test paper score and theoretical operation video score for each test in each rotation department can be obtained, and these are denoted as Q and S respectively. Therefore, the theoretical data for each rotation department can be obtained as {(Q1,S1),(Q2,S2),(Q3,S3)...(Q...S1)}. n ,S n )}, where Q k and S k The score represents any theoretical test and theoretical practical test taken during the rotation. It should be noted that clinical medical interns usually need to rotate through various departments. During the rotation, they need to take theoretical tests and theoretical practical tests related to the current department. The theoretical tests and theoretical practical tests are usually conducted simultaneously. Specifically, there will be theoretical tests and theoretical practical tests for the same knowledge point.

[0070] The system interfaces with the hospital's HIS / EMR system to collect electronic medical records written by interns in each rotation department (the electronic medical records are anonymized to remove patients' personal information), attendance schedules and actual attendance records, and clinical practice videos (4K panoramic cameras and first-person perspective cameras, such as head-mounted GoPros, are deployed in the operating room to record the surgical process synchronously, and the clinical practice videos are annotated with timestamps of key steps, such as incision, hemostasis, resection, suturing, etc.). It should be noted that the electronic medical records and clinical practice videos for each attendance are usually synchronized. Specifically, for the same attendance session, interns need to write electronic medical records and conduct clinical practice under the guidance of their supervising teachers.

[0071] It should be noted that clinical medical interns usually need to rotate through various departments, and each department will have a corresponding supervising teacher responsible for managing the interns;

[0072] The theoretical data integration module integrates the theoretical data of interns based on their rotations in different departments to obtain the theoretical data for each department. The specific method is as follows:

[0073] Step 1: The theoretical data of the interns in their rotation departments = {(Q1,S1),(Q2,S2),(Q3,S3)...(Q n ,S n )} Calculate the theoretical score F for any test (including theoretical written tests and theoretical practical tests). k The specific calculation formula is as follows:

[0074]

[0075] Where α Q α is the weighting factor for the theoretical test paper. S This is the weighting factor for the theoretical practical test; from this, the theoretical score F for each test unit in the intern's rotation department can be obtained. k ;

[0076] Step 2: Assume there is a passing score, denoted as F. 合格 And compare it with the theoretical score F of each test unit. k Comparison, if the theoretical score F of the test unit k ≥F 合格 If the intern's theoretical knowledge of the test unit meets the requirements, then the score of the theoretical test paper and the score of the theoretical video operation test for that test unit will be used as the final score; if the theoretical score of the testing unit is F... k <F 合格 If the test unit indicates that the intern's grasp of the theoretical knowledge of that test unit is insufficient and needs to be strengthened, then the test unit will be recorded as the strengthening unit.

[0077] Step 3: For the enhanced training units, interns will undergo intensive training. After the intensive training is completed, they will be re-tested with written and video exams on the enhanced units until their theoretical scores for the enhanced units are greater than or equal to the set passing score. The specific methods of intensive training are as follows:

[0078] Strengthen theoretical knowledge training: Use NLP to extract core knowledge points from medical literature and medical record databases, use graph database (Neo4j) to establish the relationship between knowledge points (symptom-treatment-surgical method-drug), and use a similarity algorithm (cosine similarity) to match the most relevant knowledge points based on the interns' theoretical test papers. Automatically extract relevant questions and answer explanations from the related knowledge points, dynamically construct a theoretical test paper to strengthen the theoretical knowledge points, and send it to the interns to help them strengthen their theoretical knowledge training.

[0079] Strengthen theoretical and practical training: Set a standard operation video corresponding to the theoretical operation test of the test unit. Usually, the standard operation video is a video recorded by the instructor performing a clinical operation simulation, or the theoretical test video with the highest video operation score in this test unit is directly selected as the standard operation video; use computer vision to identify key steps, and timestamp the key steps. Send the annotated standard operation video to the interns to help them strengthen their theoretical and practical training.

[0080] Step Four: When an intern passes all test units in their rotation department, it indicates that the intern has completed their rotation in that department and will move on to the next rotation department until all rotations are completed. If an intern passes all test units in a rotation department on their first attempt, the written test score and video operation score will be used as the final score for that rotation. If an intern does not pass all test units in a rotation department on their first attempt, the written test score and video operation score of the last passed test will be used as the final score for that rotation, and the number of times each test unit was taken will be counted. The intern's theoretical data for that rotation department will be integrated into: {(Q1,S1,C1),(Q2,S2,C2),(Q3,S3,C3)...(Q n ,S n C n )}, where C represents the number of tests; thus, the theoretical data for each rotation department throughout the entire internship period can be integrated into Where m is a positive integer, and m represents the total number of rotating departments;

[0081] By combining theoretical exams and practical tests, a comprehensive score is calculated to effectively measure interns' mastery of knowledge points in various departments. NLP technology is used to extract core knowledge points, knowledge graphs are used to analyze knowledge relationships, and personalized exams are dynamically constructed based on test results for precise reinforcement training. Based on set passing thresholds, the system automatically identifies knowledge units requiring reinforcement and dynamically adjusts training plans to ensure interns master all key knowledge points before moving to the next rotation department. Data such as interns' test scores and pass rates are recorded, comprehensively integrating data from all rotation stages to provide mentors with visualized learning progress analysis.

[0082] The clinical data integration module integrates clinical data from interns' rotations in various departments, specifically as follows:

[0083] Step 1: Extract the intern's scheduled attendance record and actual attendance record. If the actual attendance record covers the scheduled attendance record, send the electronic medical record to Step 3 and the clinical practice video to Step 4; otherwise, proceed to Step 2.

[0084] Step 2: When the actual attendance records cannot cover the attendance schedule records, the uncovered part of the attendance schedule records is extracted as the make-up work hours. The supervising supervisor arranges the make-up work time according to their own time and the make-up work hours. When the intern completes the make-up work hours, return to Step 1.

[0085] Step 3: Extract the intern's electronic medical record and use NLP to perform logical analysis to obtain three elements: temporal rationality, causal coherence, and contradiction elimination rate. These three elements are denoted as (D... 时序合理度 D 因果连贯度 D矛盾排除率 The logic score R is calculated based on the three elements, and the specific calculation formula is as follows:

[0086] R = 0.4 × D 时序合理度 +0.4×D 因果连贯度 +0.2×D 矛盾排除率 ;

[0087] Therefore, the logical reasoning score for each electronic medical record written by the intern can be denoted as (R1, R2, R3...R...). f ), where f represents the total number of attendances;

[0088] Step 3-1: The calculation process for timing rationality is as follows:

[0089] Collect a large number of real medical records and extract standard time-series paths (such as disease progression, diagnosis and treatment steps, etc.), or have the supervising teacher formulate and publish standard time-series paths; construct standard time sequences based on the standard time-series paths (such as preoperative examination → preoperative assessment → anesthesia → surgery → postoperative observation), and record the standard time sequence as the standard process sequence B (b1,b2,b3...bn); identify event timestamps in electronic medical records (such as admission, examination, surgery, medication, etc.), use dynamic time warping (DTW) to calculate the time series deviation in electronic medical records, and record it as the event flow P (p1,p2,p3...pn) of electronic medical records;

[0090] Calculate DTW distance, Where d(s) i b j ) is the distance between two events, and W is the alignment path, i.e., how to match two sequences. The optimal alignment path is obtained through dynamic path planning calculation. The specific path planning calculation formula is:

[0091]

[0092] The maximum possible DTW distance is the maximum DTW cost when two sequences are completely misaligned or unrelated, defined by the following formula: Where i and j are any events in the standard process sequence B (b1,b2,b3...bn) and the event flow P (p1,p2,p3...pn) of the electronic medical record, respectively; For event matching error;

[0093] Formula for calculating timing rationality: The smaller the deviation in the order of events, the higher the score; the smaller the score, the more events are reversed or missing.

[0094] Step 3-2: The process of calculating causal coherence is as follows:

[0095] Dependency parsing was used to analyze causal sentences in electronic medical records and extract causal relationships, including the logical chain of etiology → diagnosis → treatment → prognosis. For example, in the sentence "Patient's hypertension led to renal failure," (hypertension, renal failure) was extracted as a causal pair. Based on the NLP processing results, the causal relationships in the medical records were organized into a sequence denoted as {(G1,E1),(G2,E2),(G3,E3)...(G... n E n In the equation, G is the cause and E is the result, and G and E constitute a causal relationship.

[0096] Using a medical knowledge graph (Neo4j), standard causal logical relationships are obtained from medical guidelines or textbooks, forming a standard causal sequence denoted as {(g1,e1),(g2,e2),(g3,e3)...(g...}. n ,e n )}, where g is the standard cause and e is the standard outcome, and g and e constitute a standard causal relationship;

[0097] The similarity between medical record causal pairs and standard causal pairs is calculated using text similarity: Similarity{(G,E),(g,e)}. A matching threshold of 0.8 is set, and a match is considered successful when Similarity{(G,E),(g,e)} > 0.8. The number of successfully matched causal pairs is counted, and the number of standard causal relationships is calculated: Number of standard causal relationships = Total number of standard causal pairs related to the disease in the medical knowledge base.

[0098] Calculate causal coherence:

[0099] If the logic of the medical record conforms to medical guidelines, the causal coherence is close to 1; if the medical record contains logical errors or missing causal chains, the causal coherence is smaller.

[0100] Step 3-3: The calculation process for the conflict elimination rate is as follows:

[0101] NLP is used to identify internal contradictions in medical records, such as: "Patient has no history of diabetes" → no medical history, "Patient is taking metformin" → diabetes treatment. The contradiction is: no history of diabetes vs. diabetes treatment. The number of conflict detections in the medical record text is recorded as the actual number of conflicts, and the total number of possible conflicts is recorded as the total number of possible conflicts in all places in the medical record where contradictions may occur.

[0102] Calculate the contradiction elimination rate:

[0103] If the medical record conforms to the guideline logic, the score will be close to 1; otherwise, incorrect or missing causal relationships will result in a smaller score.

[0104] Step Four: Extract clinical practice videos and analyze the accuracy of the interns' clinical practice to obtain the two elements of practice. The two elements are: temporal alignment and gesture accuracy, denoted as (A...). 时序对齐度 A 手势精准度 The accuracy score H is calculated based on the two elements, and the specific calculation formula is as follows:

[0105] H = 0.56 × A 时序对齐度 +0.44×A 手势精准度 ;

[0106] Therefore, the accuracy score for each clinical practice video written by the intern can be recorded as (H1, H2, H3...H...). f );

[0107] Step 4-1: Analyze the timing alignment of key steps to obtain the timing alignment degree. The specific process is as follows:

[0108] Set up a standard operation video (pre-recorded surgical procedures performed by the instructor or professional doctor, with key steps marked with timestamps, such as incision, hemostasis, suturing, etc.);

[0109] Computer vision (CV) and natural language processing (NLP) were used to analyze speech commands in clinical practice videos of interns to extract the actual operation timestamps of key steps. Dynamic time warping (DTW) was employed to calculate the temporal alignment between the intern's operation and the standard operation, denoted as A. 时序对齐度 ;

[0110] Step 4-2: Analyze the gestures in key steps to obtain gesture accuracy. The specific process is as follows:

[0111] Using MediaPipeHands / OpenPose, key points of the hand (fingertips, joints, wrists, etc.) were extracted. The key coordinates at frame i of the intern's clinical practice video and the key point coordinates at frame i of the standard operation video were compared using Euclidean distance to measure the gesture deviation. The gesture deviations of each key step were then summed to obtain the total gesture deviation, denoted as V. gesture ;

[0112] The time matching error between clinical practice videos and standard operating procedure videos is calculated using dynamic time warping and denoted as V. DTW ;

[0113] Total gesture deviation V gesture Matching error V DTW The gesture accuracy A is obtained through formulaic calculation and analysis. 手势精准度 The specific calculation formula is as follows:

[0114]

[0115] Where max(gestuer) is the maximum allowed gesture error, max(DTW) is the maximum allowed time matching error, and β gesture β is the weighting factor for the total gesture deviation. DTW The weighting factor for the set time matching error;

[0116] Step 5: Integrate the clinical data (electronic medical record and clinical practice) for each attendance into {(R1,H1),(R2,H2),(R3,H3)...(R f H f The clinical data of interns throughout their rotations in various departments will be integrated into a unified whole.

[0117] By combining scheduled attendance with actual attendance records, the system automatically identifies absences and arranges make-up shifts to ensure the integrity of internship time. The logicality of electronic medical records is comprehensively evaluated based on their temporal consistency, causal coherence, and contradiction elimination rate, resulting in a logicality score. Simultaneously, clinical practice videos and standard operating procedure videos are comprehensively assessed based on temporal alignment and gesture accuracy to evaluate the accuracy of interns' clinical practice, yielding an accuracy score. This system integrates interns' clinical data from multiple dimensions, including electronic medical records, attendance records, and clinical practice videos, providing instructors with visualized learning progress analysis.

[0118] The intelligent assessment module performs an intelligent assessment of the intern's performance throughout the entire period based on integrated theoretical and clinical data. Specifically:

[0119] The integrated theoretical data The theoretical score Lm for interns in each rotation department is calculated by weighted summation. The specific calculation formula is as follows:

[0120]

[0121] Where γ Q γ S and γ C These are the weighting factors for the exam score, video operation score, and number of tests, respectively. Therefore, the theoretical scores assigned to interns in each rotation department can be denoted as (L1, L2, L3...L...). m );

[0122] Integrate the clinical data The practical score T for each intern in each rotation department is calculated by weighted summation. The specific calculation formula is as follows:

[0123]

[0124] Where γR γ H The weighting factors for logical reasoning and accuracy scores are used to assign practical scores to interns in each rotation department, which are then denoted as (T1, T2, T3...T...). m );

[0125] The internship score for each intern is calculated by weighting the theoretical score (L) and practical score (T) in each rotation department, and then visualized in tabular form. This helps supervising teachers to accurately assess the intern's learning progress throughout the entire process.

[0126] By combining the theoretical and clinical scores of interns in each rotation department for comprehensive analysis, the overall ability of interns can be evaluated in all aspects, reducing subjective assessment errors. The internship scores of interns in each rotation department are obtained, and data visualization tools are used to present the learning progress of interns at each stage. This helps supervising teachers to accurately formulate training programs, provides data support for hospitals to optimize internship management, and improves the overall teaching quality.

[0127] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A clinical medical graduation internship teaching management system, comprising a memory, a theoretical data integration module, a clinical data integration module, and an intelligent assessment module; characterized in that: The memory is used to integrate and store the theoretical and clinical data of interns in various rotation departments. The specific theoretical data includes the scores of theoretical test papers and theoretical video operation tests, while the clinical data includes the electronic medical records written by interns, the attendance schedule and actual attendance records, and clinical practice videos. The theoretical data integration module integrates theoretical data from interns' rotations in various departments. Where n is a positive integer, m is the total number of rotating departments; Q is the score of the theoretical test paper, S is the score of the theoretical video operation test, and C is the number of tests; The clinical data integration module integrates clinical data from interns' rotations in various departments. Where f is a positive integer, representing the total number of attendances, m is the total number of rotating departments; R is the logical reasoning score, and H is the accuracy score. The intelligent assessment module performs an intelligent assessment of the intern's performance throughout the entire period based on integrated theoretical and clinical data. Specifically: The integrated theoretical data is weighted and summed to obtain the theoretical score L for each intern in each rotation department. Therefore, the theoretical scores for each intern in each rotation department are denoted as (L1, L2, L3...L...). m ); The integrated clinical data is weighted and summed to obtain the internship scores for each intern in each rotation department. Therefore, the internship scores for each intern in each rotation department are denoted as (T1, T2, T3...T...). m ); The internship score for each intern is calculated by weighting the theoretical score (L) and practical score (T) in each rotation department, and then visualized in tabular form.

2. The clinical medical graduation internship teaching management system according to claim 1, characterized in that, The specific methods for integrating theoretical and clinical data are as follows: The theoretical test papers and theoretical practical test videos of interns in each rotation department were collected. The supervising teachers scored the theoretical test papers and videos separately to give scores. This yielded the theoretical test paper score and theoretical practical test video score for each test in each rotation department, which were denoted as Q and S respectively. Therefore, the theoretical data for each rotation department can be calculated as {(Q1,S1),(Q2,S2),(Q3,S3)...(Q...S1)}. n ,S n )}, where Q k and S k The score is the score of any one theoretical test and theoretical practical test in the rotating department; It connects with the hospital's HIS / EMR system to collect electronic medical records written by interns in various rotation departments, their scheduled attendance records and actual attendance records, and clinical practice videos, and records them as clinical data.

3. The clinical medical graduation internship teaching management system according to claim 1, characterized in that, The specific methods for interns to integrate theoretical data from their rotations in various departments are as follows: Step 1: The theoretical data of the interns in their rotation departments = {(Q1,S1),(Q2,S2),(Q3,S3)...(Q n ,S n )}, calculate the theoretical score F for any given test. k The specific calculation formula is as follows: Where α Q α is the weighting factor for the theoretical test paper. S This is the weighting factor for the theoretical practical test; from this, the theoretical score F for each test unit in the intern's rotation department can be obtained. k ; Step 2: Assume there is a passing score, denoted as F. 合格 And compare it with the theoretical score F of each test unit. k Compare; if the theoretical score F of the test unit k <F 合格 When this happens, the test unit is recorded as a reinforcement unit; Step 3: For the enhanced units, implement intensive training for interns. After the intensive training is completed, conduct the theoretical paper test and theoretical video test for the enhanced units again until the theoretical score of the enhanced units is greater than or equal to the set passing score. Step Four: When an intern passes all test units in their rotation department, they proceed to the next rotation department until all rotations are completed. If an intern passes all test units in a rotation department on their first attempt, the written test score and video operation score are used as the final score for that rotation. If an intern fails all test units in a rotation department on their first attempt, the written test score and video operation score of the last passed test are used as the final score for that rotation, and the number of tests for that unit is counted. The intern's theoretical data for that rotation department is integrated into: {(Q1,S1,C1),(Q2,S2,C2),(Q3,S3,C3)...(Q n ,S n C n )}, where C represents the number of tests; thus, the theoretical data for each rotation department throughout the entire internship period can be integrated into Where m is a positive integer, and m represents the total number of rotating departments.

4. The clinical medical graduation internship teaching management system according to claim 3, characterized in that, The specific methods of intensive training are as follows: Strengthen theoretical knowledge training: Use NLP to extract core knowledge points from medical literature and medical record databases, use graph databases to establish relationships between knowledge points, use similarity algorithms to match the most relevant knowledge points based on the interns' theoretical test scores, automatically extract relevant questions and answer explanations from the relevant knowledge points, dynamically construct a theoretical test to strengthen theoretical knowledge, and send it to the interns. Strengthen theoretical and practical training: Set up a standard operation video corresponding to the theoretical operation test of the test unit, use computer vision to identify key steps, and timestamp the key steps. Then send the annotated standard operation video to the interns.

5. The clinical medical graduation internship teaching management system according to claim 1, characterized in that, The specific methods for integrating clinical data are as follows: Step 1: Extract the intern's scheduled attendance record and actual attendance record. If the actual attendance record covers the scheduled attendance record, send the electronic medical record to Step 3 and the clinical practice video to Step 4; otherwise, proceed to Step 2. Step 2: When the actual attendance records cannot cover the attendance schedule records, the uncovered part of the attendance schedule records is extracted as the make-up work hours. The supervising supervisor arranges the make-up work time according to their own time and the make-up work hours. When the intern completes the make-up work hours, return to Step 1. Step 3: Extract the intern's electronic medical record and use NLP to perform logical analysis to obtain three elements: temporal rationality, causal coherence, and contradiction elimination rate. These three elements are denoted as (D... 时序合理度 D 因果连贯度 D 矛盾排除率 The logical reasoning score R is calculated based on the three elements. Therefore, the logical reasoning score for each electronic medical record written by the intern is denoted as (R1, R2, R3...R...). f ), where f represents the total number of attendances; Step Four: Extract clinical practice videos and analyze the accuracy of the interns' clinical practice to obtain the two elements of practice. The two elements are: temporal alignment and gesture accuracy, denoted as (A...). 时序对齐度 A 手势精准度 The accuracy score H is calculated based on the two elements, and the accuracy score of each clinical practice video written by the intern is recorded as (H1, H2, H3...H). f ); Step 5: Integrate the clinical data from each attendance session into {(R1,H1),(R2,H2),(R3,H3)...(R f H f The clinical data of interns throughout their rotations in various departments will be integrated into a unified whole.

6. The clinical medical graduation internship teaching management system according to claim 5, characterized in that, The logical evaluation method for electronic medical records is as follows: Step 3-1: The calculation process for timing rationality is as follows: Collect a large number of real medical records and extract standard time sequence paths, or have the supervising teacher formulate and publish standard time sequence paths; construct a standard time sequence based on the standard time sequence paths, and record the standard time sequence as a standard process sequence B(b1,b2,b3...bn); identify event timestamps in electronic medical records, use dynamic time warping to calculate the time sequence deviation in electronic medical records, and record it as the event flow P(p1,p2,p3...pn) of electronic medical records; The formula for calculating DTW distance and maximum possible DTW distance, and the formula for calculating time series reasonableness are as follows: Step 3-2: The process of calculating causal coherence is as follows: Dependency parsing was used to analyze causal sentences in electronic medical records and extract causal relationships, including the logical chain of etiology → diagnosis → treatment → prognosis. Based on the NLP processing results, the causal relationships in the medical records were organized into a sequence denoted as {(G1,E1),(G2,E2),(G3,E3)...(G... n E n In the equation, G is the cause and E is the result, and G and E constitute a causal relationship. Using medical knowledge maps, standard causal logical relationships are obtained from medical guidelines or textbooks, forming a standard causal sequence denoted as {(g1,e1),(g2,e2),(g3,e3)...(g...}. n ,e n )}, where g is the standard cause and e is the standard outcome, and g and e constitute a standard causal relationship; The similarity between medical record causal pairs and standard causal pairs is calculated using text similarity: Similarity{(G,E),(g,e)}. A matching threshold of 0.8 is set, and a match is considered successful when Similarity{(G,E),(g,e)} > 0.

8. The number of successfully matched causal pairs is counted, and the number of standard causal relationships is calculated: Number of standard causal relationships = Total number of standard causal pairs related to the disease in the medical knowledge base. Calculate causal coherence: Step 3-3: The calculation process for the conflict elimination rate is as follows: NLP is used to identify contradictions within medical records. The number of conflict detections in the medical record text is recorded as the actual number of conflicts. The total number of possible conflicts is calculated as the number of all possible conflict pairs that can be detected in the medical record. Calculate the contradiction elimination rate:

7. A clinical medical graduation internship teaching management system according to claim 6, characterized in that, The DTW distance and the maximum possible DTW distance are calculated as follows: Where d(s) i ,b j ) is the distance between two events, and W is the alignment path, i.e., how to match two sequences. The optimal alignment path is obtained through dynamic path planning calculation. The specific path planning calculation formula is: The maximum possible DTW distance is the maximum DTW cost when two sequences are completely misaligned or unrelated, defined by the following formula: Where i and j are any events in the standard process sequence B (b1,b2,b3...bn) and the event flow P (p1,p2,p3...pn) of the electronic medical record, respectively; This represents the event matching error.

8. A clinical medical graduation internship teaching management system according to claim 5, characterized in that, The accuracy assessment method for clinical practice videos is as follows: Step 4-1: Analyze the timing alignment of key steps to obtain the timing alignment degree. The specific process is as follows: A standard operation video is set up. Computer vision and NLP are used to analyze the voice commands in the intern's clinical practice video to extract the actual operation timestamps of key steps. Dynamic time warping is used to calculate the timing alignment between the intern's operation and the standard operation, denoted as A. 时序对齐度 ; Step 4-2: Analyze the gestures in key steps to obtain gesture accuracy. The specific process is as follows: Hand key points were extracted using MediaPipe Hands / OpenPose. The key coordinates at frame i of the intern's clinical practice video and the key point coordinates at frame i of the standard operation video were compared using Euclidean distance to measure the gesture deviation. The gesture deviations of each key step were then summed to obtain the total gesture deviation, denoted as V. gesture ; The time matching error between clinical practice videos and standard operating procedure videos is calculated using dynamic time warping and denoted as V. DTW ; Total gesture deviation V gesture Matching error V DTW The gesture accuracy A is obtained through formulaic calculation and analysis. 手势精准度 The specific calculation formula is as follows: Where max(gestuer) is the maximum allowed gesture error, max(DTW) is the maximum allowed time matching error, and β gesture β is the weighting factor for the total gesture deviation. DTW The weighting factor for the time matching error is set.

Images