Clinical Medicine Practice Skills Assessment System

Abstract

The present invention discloses a clinical medicine practice skills assessment system, which relates to the field of clinical medicine education technology. The system includes the following components: a comprehensive data processing module: collecting and preprocessing the electronic medical record data of real patients and the multimodal data of candidates, and providing data for subsequent modules by fusing the characteristics of each modal data and cleaning the electronic medical record data; the present invention generates personalized intelligent virtual patients based on the electronic medical record data of real patients, providing candidates with rich and diverse clinical simulation scenarios, which not only have the complexity and diversity of real patients, but also can adjust the condition in real time according to the decision of the candidates, simulate the uncertainty and complexity of clinical scenarios, and help candidates cultivate and improve clinical thinking in practice, so that they learn to make accurate judgments and decisions when the condition changes. At the same time, through interaction with virtual patients, candidates can also exercise their adaptability and better adapt to various challenges in actual clinical work.

Description

Technical Field

[0001] The present invention relates to the technical field of clinical medicine education, and in particular to a clinical medicine practice skills assessment system. Background Art

[0002] In the field of clinical medical education and assessment, the assessment of practical skills has always been a key point and difficulty. With the continuous advancement of medical technology and the growing demand for medical care, the requirements for medical students' clinical practical skills are becoming higher and higher. The traditional medical education model has gradually exposed its limitations in cultivating practical skills. Therefore, it is particularly important to seek an efficient, comprehensive and safe method for assessing clinical practical skills. In recent years, with the rapid development of intelligent technology and virtual reality technology, new ideas and means have been provided for the assessment of clinical medical practical skills.

[0003] However, traditional clinical medicine assessment methods have many shortcomings. Although theoretical exams can test students' mastery of medical knowledge, they are limited to the memory of fixed knowledge and it is difficult to truly reflect students' ability to make accurate judgments and decisions based on changes in the disease in actual clinical scenarios. Although clinical internships provide students with certain practical opportunities, they are limited by the uncontrollability and diversity of real cases. Students have limited opportunities to come into contact with typical complex cases, and there are risks in performing operations on real patients. In addition, some existing simulation assessment systems lack personalization and cannot generate targeted simulation cases based on real patient data. The evaluation method is single and is only based on single-dimensional data. It ignores the examinee's posture, voice communication, physiological indicators, and images and videos during the operation, as well as multimodal information such as the operation process. As a result, the evaluation results are incomplete and inaccurate, and it is difficult to truly reflect the examinee's clinical practice skills.

[0004] In response to the above problems, it is necessary to optimize the existing clinical medicine practice skills assessment system by using the electronic medical record data of real patients to generate personalized intelligent virtual patients, simulate changes in the disease in real time, and integrate multimodal data for comprehensive evaluation. Therefore, it is of great significance to develop a clinical medicine practice skills assessment system that can comprehensively realize the above characteristics. Summary of the Invention

[0005] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide a clinical medicine practice skills assessment system. It can generate personalized intelligent virtual patients by using the electronic medical record data of real patients, provide candidates with a variety of clinical simulation scenarios, and adjust the condition in real time according to the candidate's decision through the dynamic condition control module, simulating the uncertainty and complexity of the clinical scenario. The comprehensive data processing module integrates multimodal data for evaluation, and comprehensively reflects the candidate's clinical practice skills from multiple angles. The assessment and analysis module uses a multi-dimensional evaluation index system and combines expert experience for comprehensive evaluation, providing a scientific and objective basis for clinical medicine education and assessment. The interactive display module uses virtual reality or augmented reality technology to provide candidates with a more realistic clinical simulation experience, thereby improving the candidate's participation and learning effect.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a clinical medicine practice skills assessment system, which includes the following components:

[0007] Comprehensive data processing module: collects and preprocesses real patient electronic medical record data and examinee multimodal data, integrates the features of each modality data, cleans the electronic medical record data, and provides data for subsequent modules;

[0008] Intelligent virtual patient generation module: This module analyzes pre-processed electronic medical record data and builds an intelligent virtual patient model by combining typical electronic medical record features, random features, medical knowledge graph features, and new medical research features. It also introduces uncertainty in the condition to generate a virtual patient condition that conforms to clinical reality.

[0009] Dynamic disease control module: monitors the candidates' decisions in real time during the simulation assessment, calculates the change in the virtual patient's condition based on the decision, and predicts the evolution of the condition, so that the virtual patient's condition changes dynamically with the candidate's decision, simulating the real clinical diagnosis and treatment process;

[0010] Assessment and Analysis Module: This module constructs a multi-dimensional assessment indicator system, calculates the examinee's scores on each individual assessment indicator and the overall score, and comprehensively evaluates the examinee's clinical practice skills by adjusting and optimizing the indicator weights;

[0011] Interactive display module: Use virtual reality or augmented reality technology to design an interactive interface, achieve natural interaction through voice and gesture recognition technology, and display the patient's condition changes, examination results and operation prompts in real time.

[0012] Furthermore, the comprehensive data processing module collects and preprocesses the electronic medical record data of real patients and the multimodal data of candidates. The electronic medical record data includes the patient's basic information, medical history, symptoms and various examination indicators, and the multimodal data of candidates includes movement and posture data, voice communication data, physiological indicator data, and image and video data.

[0013] Furthermore, the intelligent virtual patient generation module parses the pre-processed electronic medical record data. Specifically, it receives the electronic medical record data pre-processed by the comprehensive data processing module, parses the data format, extracts key information related to disease diagnosis and treatment and stores it in a database, conducts in-depth analysis and learning on the parsed data, explores the hidden rules and patterns in the data, and builds an intelligent virtual patient model based on it. At the same time, in order to make the virtual patient close to the real clinical situation, uncertainty factors are introduced into the model, so that the manifestation and development of each virtual patient's condition are random and different.

[0014] Furthermore, the intelligent virtual patient generation module constructs an intelligent virtual patient model, and its model formula is: P = α·E+β·R+γ·N+δ·C, wherein P is the condition feature vector of the generated intelligent virtual patient, which contains various condition information of the patient, E is a typical condition feature vector extracted from electronic medical record data, R is a randomly generated condition feature vector, N is a condition feature vector generated based on a medical knowledge graph, reflecting the general laws of the disease, and C is a condition feature vector generated based on current medical research hotspots and new discoveries, making the virtual patient more timely. α, β, γ, and δ are weight coefficients of E, R, N, and C, respectively. Different factors are scored and determined according to their importance in the generation of virtual patients. Different diseases and evaluation scenarios have different degrees of dependence on each factor. Based on the understanding of the disease, it is judged which factors are more critical to the generation of the virtual patient's condition, thereby determining the weights.

[0015] Furthermore, the intelligent virtual patient generation module introduces uncertainty factors into the model, and its uncertainty adjustment formula is: Among them, ΔP is the introduced uncertainty adjustment amount of the disease, which is used to increase the uncertainty in the disease generation process. U is the uncertainty coefficient, which is set according to the actual clinical situation and evaluation requirements to control the size of the uncertainty. i is the i-th eigenvalue in the disease characteristic vector, is the mean of all eigenvalues ​​in the disease feature vector, and m is the number of dimensions of the disease feature vector.

[0016] Furthermore, the dynamic disease control module monitors the decisions of the examinees in the simulation assessment in real time. Specifically, it monitors the diagnosis and treatment decisions of the examinees in the simulation assessment process in real time, compares and analyzes the decisions with the pre-set standard diagnosis and treatment plans, judges the rationality and effectiveness of the decisions, and establishes a disease evolution model that can reflect the development laws of the disease and the treatment effects. It uses the examinee's decision and the current condition of the virtual patient as input to predict the development direction and degree of change of the disease. According to the prediction results of the disease evolution model, it adjusts the condition of the intelligent virtual patient in real time, and dynamically displays the development and changes of the disease by changing the symptoms and examination indicators of the virtual patient.

[0017] Furthermore, the disease dynamic control module establishes a disease evolution model that can reflect the disease development law and treatment effect, and the model formula is: Among them, S t+1 is the predicted disease status value of the next time step, S t is the disease state value at the current time step, θ is the disease evolution coefficient, which controls the evolution speed of the disease with decision and time, and ΔS is the change in disease caused by the current decision. Its calculation formula is: ΔS is the change in the condition of the virtual patient after the candidate makes a decision, l is the number of decisions taken by the candidate, which is determined by the number of decisions actually taken by the candidate during the simulation assessment process, and a j is the effect coefficient of the jth decision, reflecting the impact of the decision on the condition, D j is the intensity factor of the jth decision, which is determined according to the specific content of the candidate's decision, b j is the decision sensitivity coefficient, which controls the response speed of the disease to the decision, S0 is the current disease state value of the virtual patient, c j is the threshold of the j-th decision, is the random interference coefficient, which introduces randomness to simulate clinical uncertainty, ∈ is the random interference term, and obeys the probability distribution.

[0018] Furthermore, the assessment and analysis module constructs an evaluation index system covering multiple dimensions, including evaluation indicators of diagnostic accuracy, rationality of treatment plan, emergency response ability, communication ability, operational standardization, concentration and psychological state.

[0019] Furthermore, the assessment analysis module calculates the examinee's score on each individual assessment indicator and the comprehensive score, and comprehensively evaluates the examinee's clinical practice skill level by adjusting and optimizing the indicator weights. The score calculation formula for each individual assessment indicator is: Among them, Score single is the score of the examinee on a single assessment indicator, p is the number of sub-assessment items under the single indicator, is the weight of the ith sub-evaluation item, which is determined according to the importance of each sub-evaluation item in the single indicator. is the candidate's matching score on the ith sub-assessment item, which is obtained by comparing the candidate's performance on this sub-assessment item with the standard. The comprehensive score formula is: Among them, Score total is the comprehensive score of the examinee, reflecting the examinee's clinical practice skill level, q is the total number of evaluation indicators, ω k is the weight of the k-th evaluation metric, is the examinee’s score on the kth evaluation indicator.

[0020] Compared with the existing technology, this clinical medicine practice skills assessment system has the following beneficial effects:

[0021] 1. The present invention generates personalized intelligent virtual patients based on the electronic medical record data of real patients, providing candidates with a rich variety of clinical simulation scenarios. These scenarios not only have the complexity and diversity of real patients, but can also adjust the condition in real time according to the candidate's decision, simulating the uncertainty and complexity of clinical scenarios. This helps candidates cultivate and improve their clinical thinking in practice, enabling them to learn to make accurate judgments and decisions when the condition changes. At the same time, through interaction with virtual patients, candidates can also exercise their adaptability and better adapt to the various challenges in actual clinical work.

[0022] 2. The present invention can integrate multimodal data for evaluation through a comprehensive data processing module, including multi-dimensional information such as movement posture, voice communication, physiological indicators, and images and videos during the operation. This comprehensive evaluation method can more realistically reflect the clinical practice skill level of the examinee, avoiding the one-sidedness and inaccuracy that may be brought about by a single-dimensional evaluation. In addition, the assessment and analysis module uses a multi-dimensional evaluation index system and combines expert experience for comprehensive evaluation, which further improves the scientificity and objectivity of the evaluation results. It not only helps to accurately measure the skill level of the examinee, but also provides a strong basis for clinical medical education and assessment, and promotes the sustainable development of clinical medical education.

[0023] Other advantages, objects and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be learned from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0025] Figure 1 This is the process operation diagram of the clinical medicine practice skills assessment system;

[0026] Figure 2 This is a flowchart of the clinical medicine practice skills assessment system. DETAILED DESCRIPTION

[0027] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0028] Example 1

[0029] In the clinical practice skills assessment scenario of medical schools, the clinical medicine practice skills intelligent assessment system of the present invention plays an important role.

[0030] The school's teaching hospital provides electronic medical record data of real patients through standardized interfaces, covering rich case information from multiple departments such as internal medicine, surgery, obstetrics and gynecology, and pediatrics. At the same time, in simulated assessment scenarios, high-precision motion capture equipment is used to record the posture information of candidates during surgical operations, physical examinations, and other actions at a frequency of 120 frames per second to ensure that every subtle movement can be accurately captured. The microphone array collects all-round voice communication content between candidates and virtual patients and team members, and has excellent noise reduction performance. It can clearly collect voice data even in noisy simulated environments. The physiological monitor monitors the candidate's heart rate, EEG, blood pressure and other physiological indicators in real time. The camera records image and video data during the operation at 4K resolution. The system preprocesses these data and performs multimodal data fusion.

[0031] According to the teaching syllabus and assessment requirements, the system extracts typical disease characteristics from electronic medical record data, combines randomly generated features, knowledge in the medical knowledge graph, and the latest medical research results to generate intelligent virtual patients. The virtual patient disease generation formula is: P = α·E+β·R+γ·N+δ·C, where P is the disease characteristic vector of the generated intelligent virtual patient, E is the typical disease characteristic vector extracted from the electronic medical record data, obtained through dimensionality reduction algorithms such as cluster analysis, R is a randomly generated disease characteristic vector, which is randomly generated by the system according to a certain probability distribution, and N is the disease characteristic vector generated based on the medical knowledge graph. The feature vector is obtained by extracting relevant information from the medical knowledge graph database. C is the disease feature vector generated based on current medical research hotspots and new discoveries, which is obtained by collecting and analyzing the latest medical research literature. α, β, γ, and δ are weight coefficients, and α+β+γ+δ=1. When generating virtual patients for appendicitis cases, clinical medical experts score the importance of different factors in the generation of virtual patients and determine α=0.5, β=0.2, γ=0.2, and δ=0.1, which means that virtual patients are generated based on typical disease characteristics in electronic medical record data. When generating 100 virtual patients with appendicitis, the clinical medical experts score the importance of different factors in the generation of virtual patients and determine α=0.5, β=0.2, γ=0.2, and δ=0.1, which means that virtual patients are generated based on typical disease characteristics in electronic medical record data. After the system was used for simulation assessment, the feedback from candidates and expert evaluation opinions were collected. It was found that 30% of the candidates reported that the virtual patient's condition was too stereotyped and lacked variation. Further analysis found that the randomly generated features did not contribute enough to the diversity of the condition. The weights were adjusted to α = 0.4, β = 0.3, γ = 0.2, δ = 0.1, and 100 virtual patients were generated again for assessment. The candidates reported that the diversity of the condition had improved, and the proportion of candidates who believed that the condition was more in line with the actual situation increased to 70%. For the virtual patient with appendicitis, the system simulated different stages of the disease, such as the atypical early stage, the typical right lower abdominal pain stage, and the stage of symptom stagnation. When generating disease characteristics, not only are symptom manifestations (such as the nature, degree, and location changes of abdominal pain, and accompanying symptoms such as nausea and vomiting) considered, but laboratory test indicators (such as white blood cell count and neutrophil ratio) and imaging test results (such as the shape, size, and surrounding exudation of the appendix) are also combined. At the same time, based on new advances in medical research, such as the impact of minimally invasive treatment methods for appendicitis on the disease, the disease characteristics and treatment response of the virtual patient are adjusted. In order to make the generated disease more consistent with actual uncertainty, the disease uncertainty formula is introduced: Among them, ΔP is the introduced uncertainty adjustment for the disease condition, and U is the uncertainty coefficient. Based on the actual clinical situation and assessment requirements, U = 0.3 was initially set. During the assessment, it was found that candidates generally reflected that the changes in the disease condition were highly predictable and the assessment difficulty was relatively low. After analyzing the assessment results of 80 candidates, it was found that the average score was high and the discrimination was not obvious. Therefore, the U value was adjusted to 0.5. After the assessment was conducted again, the average score of the candidates decreased, and the standard deviation of the score increased from 10 to 15, indicating that the discrimination was improved. At this time, it is considered that U = 0.5 is more appropriate, P i is the i-th eigenvalue in the disease characteristic vector, is the mean of all eigenvalues ​​in the disease feature vector, and m is the number of dimensions of the disease feature vector, which is determined by the construction method of the disease feature vector.

[0032] During the exam, the system monitors the candidate's diagnosis and treatment plan in real time. When the candidate makes decisions such as issuing a checkup order and formulating a treatment plan, the system calculates the change in the condition based on factors such as the effectiveness and intensity of the decision, and predicts the progression of the condition. For example, if a candidate is diagnosed with appendicitis and chooses to use antibiotics for treatment, the system will determine the progression of the condition based on information such as the type and dosage of the antibiotics and the medication time entered by the candidate. If the drug is selected correctly and the dosage is appropriate, the virtual patient's abdominal pain symptoms will gradually subside, and indicators such as white blood cell count will gradually return to normal. If the candidate delays diagnosis or the treatment plan is unreasonable, such as not undergoing surgery in time, resulting in a perforated appendix, the system will predict the progression of the condition based on the following formula: Simulate the process of disease progression, where S t+1 is the predicted disease status value of the next time step, S t is the disease state value at the current time step, θ is the disease evolution coefficient, and θ=0.4 is initially set based on the general development law of the disease and experience. For the appendicitis disease evolution coefficient adjustment process and data, it was found in the simulation evaluation that the disease evolution speed did not match the actual situation and was too fast. After analyzing the disease evolution of 70 simulation evaluations, it was found that the actual disease evolution speed was 20% faster than the predicted speed. Therefore, the θ value was adjusted to 0.3. When the disease evolution was re-evaluated, the degree of conformity with the actual situation increased from 60% to 75%. ΔS is the change in the disease caused by the current decision. is the random interference coefficient, which is initially set according to the degree of uncertainty in actual clinical practice In the exam, it was found that the uncertainty of the disease was not well reflected. The candidates were easy to predict the development of the disease. After analyzing the feedback of 90 candidates, it was found that the proportion of candidates who believed that the disease was predictable was as high as 60%. The value was adjusted to 0.3. When the test was reassessed, the proportion of candidates who believed that the condition was predictable dropped to 40%. More appropriately, ε is a random interference term, which is randomly generated by the system according to a preset probability distribution. At this time, symptoms of peritonitis will appear, such as general abdominal pain, rebound pain, abdominal muscle tension, etc. At the same time, related examination indicators will further deteriorate. After the assessment, the system calculates the examinee's scores on individual indicators such as diagnostic accuracy, rationality of treatment plan, and communication ability based on the pre-built evaluation indicator system, and then combines the scores of various indicators to obtain the final assessment score. The single indicator scoring formula is: Among them, Score single is the score of a single indicator, and p is the number of sub-evaluation items under the single indicator, which is determined by the evaluation system developer based on the specific settings of the evaluation indicator. is the weight of the ith sub-evaluation item. In the diagnostic accuracy index, three sub-evaluation items are set: symptom judgment, disease classification, and differential diagnosis. The symptom judgment weight is determined based on the importance of each sub-evaluation item in the single indicator. Disease classification weight Differential diagnosis weight When analyzing the diagnostic accuracy index scores of 120 candidates, it was found that the scores of the disease classification sub-assessment items were generally high and had a great influence on the final score, while the scores of the differential diagnosis sub-assessment items were not obvious, so the weight was adjusted to the symptom judgment weight. Disease classification weight Differential diagnosis weight After re-evaluation, the score discrimination of the diagnostic accuracy index improved, and the standard deviation increased from 8 to 12. is the candidate's match score on the ith sub-assessment item, obtained by comparing the candidate's performance with the standard, using the comprehensive score formula: Among them, Score total is the comprehensive score, q is the total number of evaluation indicators, which is pre-set by the evaluation system developer, ω k is the weight of the kth evaluation indicator. When analyzing the results of multiple rounds of evaluation, it was found that the correlation between the score of the treatment plan rationality indicator and the actual clinical ability of the examinee was not as good as expected, while communication ability is becoming more and more important in actual clinical scenarios. Therefore, the weights were adjusted to diagnostic accuracy weight ω1 = 0.4, treatment plan rationality weight ω2 = 0.3, and communication ability weight ω3 = 0.3. After the adjustment, the evaluation results of 100 examinees were analyzed and it was found that the correlation between the comprehensive score and the examinee's performance in actual clinical practice increased from the previous 0.6 to 0.7. It is the score of the examinee on the kth evaluation indicator. When evaluating the accuracy of diagnosis, it not only considers whether the examinee correctly diagnoses the disease, but also the timeliness and comprehensiveness of the diagnosis, such as whether important symptoms or examination results are missed. When evaluating the rationality of the treatment plan, factors such as the choice of treatment method, the use of drugs, and the feasibility of surgical operations will be considered.

[0033] Candidates interact with virtual patients using virtual reality or augmented reality devices. The system arranges the order of information presentation based on urgency, relevance, timeliness, and importance. For example, if the virtual patient's condition suddenly worsens, emergency prompts are displayed to the candidate first, ensuring that the candidate can make the right decision in a timely manner. During the interaction, virtual reality devices provide a highly immersive experience for candidates. Candidates can perform detailed physical examinations in the virtual environment, such as feeling the virtual patient's muscle tension and tenderness when palpating the abdomen. Augmented reality devices can overlay relevant information about the virtual patient on the real scene. For example, during surgical operations, the device can display the anatomical structure of the surgical site, the location of important blood vessels and nerves in real time, and other information to assist candidates in performing accurate operations. At the same time, the system will update the displayed information in real time based on the candidate's operations and changes in the condition, such as examination result reports and prompts of changes in the condition, helping candidates to promptly understand the progress of the disease and the effectiveness of their own operations. This interactive method not only improves the efficiency of the assessment but also provides candidates with a more realistic clinical practice experience.

[0034] Example 2

[0035] The system of the present invention can also be effectively applied in scenarios where hospitals provide skills improvement training and regular evaluations for on-the-job doctors.

[0036] The hospital's information system provides a wealth of electronic medical record data, covering a variety of common and rare diseases. It not only includes the patient's basic diagnosis and treatment information, but also integrates multi-dimensional information such as long-term follow-up data and genetic test results. During the training and evaluation process, multimodal data is collected when doctors interact with virtual patients. Motion capture equipment can track the doctor's gestures and body postures in real time when performing various operations, such as cardiologists performing cardiac auscultation and dermatologists observing skin lesions. Voice acquisition equipment can not only clearly record the conversation between the doctor and the virtual patient, but also analyze the emotional tendencies and the use of professional terms in the voice. Physiological indicator monitoring equipment can continuously monitor the doctor's concentration and other physiological reactions during the diagnosis and treatment process, such as heart rate variability and changes in EEG activity.

[0037] The system customizes personalized virtual patients based on training objectives and evaluation requirements. For emerging diseases or treatments, the system will generate corresponding condition descriptions based on the latest medical research hotspots. For example, during the COVID-19 pandemic, the system can generate virtual patients with COVID-19 symptoms and disease progression characteristics for doctors to simulate diagnosis and treatment. During the training process, doctors interact with virtual patients multiple times, and the system continuously adjusts the condition descriptions and simulates different clinical situations to help doctors improve their ability to deal with complex conditions.

[0038] For virtual COVID-19 patients, the system will simulate different stages from asymptomatic infection, mild symptoms, severe symptoms to critical illness, as well as various possible complications, such as respiratory failure, heart damage, thrombosis, etc. At the same time, combined with the latest diagnosis and treatment guidelines, the virtual patient's response to different treatment measures will be adjusted, such as antiviral drugs, immunotherapy, respiratory support, etc. During the interaction between doctors and virtual patients, the system will provide real-time feedback on changes in the condition based on the doctor's decision, guide the doctor to think deeply and analyze the condition, and improve the level of diagnosis and treatment. In addition to emerging diseases, for some rare diseases, the system will also refer to case reports and research results around the world to generate representative virtual patients. During the training process, different difficulty levels will be set, from basic diagnosis and treatment process training to multidisciplinary collaborative treatment of complex diseases, to gradually improve the doctor's capabilities.

[0039] The system monitors doctors' decisions in real time and dynamically adjusts the virtual patient's condition based on the evolution prediction formula. After each interaction, the system immediately provides a detailed evaluation report, including the doctor's score and analysis on various evaluation indicators. For example, it points out key symptoms that the doctor missed during the diagnosis process and unreasonable aspects of the treatment plan. Doctors can improve their clinical skills in a targeted manner based on this feedback information. At the same time, hospitals can develop personalized training plans for doctors based on their comprehensive scores to improve overall medical standards.

[0040] The evaluation report not only includes the doctor's scores on traditional indicators such as diagnostic accuracy, rationality of treatment plans, and communication skills, but also evaluates the doctor's decision-making ability, teamwork ability, and ability to apply the latest medical knowledge when facing uncertainty. For example, when dealing with complex conditions, whether the doctor can communicate and collaborate with doctors in other departments in a timely manner, and whether he can accurately apply the latest research results and treatment methods. For doctors with unsatisfactory evaluation results, the hospital will arrange special mentors for one-on-one guidance and conduct intensive training on the doctor's weak links. For doctors with outstanding performance, higher-level training opportunities will be provided, such as participation in international academic exchanges, cutting-edge research projects, etc., to encourage doctors to continuously improve their professional level.

[0041] With the continuous updating of medical knowledge and the accumulation of clinical practice, the system will regularly adjust the parameters and weights of each module. For example, when new treatment guidelines are released, the system will adjust the decision-making effect coefficient and evaluation indicator weight accordingly to ensure that the content of training and evaluation is always consistent with the latest medical standards. Through this continuous optimization method, the system can continuously adapt to changes in clinical practice and provide doctors with more effective skill improvement and evaluation services.

[0042] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as above in terms of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can, without departing from the scope of the technical solution of the present invention, make some changes or modifications to equivalent embodiments using the technical contents disclosed above. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. Clinical medicine practice skills assessment system, characterized by: The system includes the following components: Comprehensive data processing module: This module collects and preprocesses real patient electronic medical record data and examinee multimodal data. Electronic medical record data includes basic patient information, medical history, symptoms, and various examination indicators. Examinee multimodal data includes posture data, voice communication data, physiological indicator data, and image and video data. By fusing the features of each modality and cleaning the electronic medical record data, it provides data for subsequent modules. Intelligent virtual patient generation module: parses the preprocessed electronic medical record data and constructs an intelligent virtual patient model by combining typical features of electronic medical records, random features, medical knowledge graph features, and new features of medical research. The model formula is: P = α·E+β·R+γ·N+δ·C, where P is the generated intelligent virtual patient's condition feature vector, which contains various condition information of the patient, E is the typical condition feature vector extracted from the electronic medical record data, R is a randomly generated condition feature vector, N is a condition feature vector generated based on the medical knowledge graph, which reflects the general laws of the disease, and C is a condition feature vector generated based on current medical research hotspots and new discoveries, making the virtual patient more timely. α, β, γ, and δ are the weight coefficients of E, R, n, and C, respectively, and condition uncertainty is introduced. The uncertainty adjustment formula is: Among them, ΔP is the uncertainty adjustment amount introduced to increase uncertainty in the disease generation process, U is the uncertainty coefficient, which controls the size of uncertainty, and P i is the i-th eigenvalue in the disease characteristic vector, is the mean of all eigenvalues ​​in the disease feature vector, m is the number of dimensions of the disease feature vector, and generates a virtual patient condition that conforms to clinical reality; Dynamic disease control module: monitors the candidates' decisions in real time during the simulation assessment, calculates the change in the virtual patient's condition based on the decision, and predicts the evolution of the condition, so that the virtual patient's condition changes dynamically with the candidate's decision, simulating the real clinical diagnosis and treatment process; Assessment and Analysis Module: This module constructs a multi-dimensional assessment indicator system, calculates the examinee's scores on each individual assessment indicator and the overall score, and comprehensively evaluates the examinee's clinical practice skills by adjusting and optimizing the indicator weights; Interactive display module: Use virtual reality or augmented reality technology to design an interactive interface, achieve natural interaction through voice and gesture recognition technology, and display the patient's condition changes, examination results and operation prompts in real time.

2. The clinical medicine practice skills assessment system according to claim 1, characterized in that: The intelligent virtual patient generation module parses the pre-processed electronic medical record data. Specifically, it receives the electronic medical record data pre-processed by the comprehensive data processing module, parses the data format, extracts key information related to disease diagnosis and treatment and stores it in a database, conducts in-depth analysis and learning on the parsed data, explores the hidden rules and patterns in the data, and builds an intelligent virtual patient model based on it. At the same time, in order to make the virtual patient close to the real clinical situation, uncertainty factors are introduced into the model so that the manifestation and development of each virtual patient's condition are random and different.

3. The clinical medicine practice skills assessment system according to claim 1, characterized in that: The dynamic disease control module monitors the decisions of the examinees in the simulation assessment in real time. Specifically, it monitors the diagnosis and treatment decisions of the examinees in the simulation assessment process in real time, compares and analyzes the decisions with the pre-set standard diagnosis and treatment plans, judges the rationality and effectiveness of the decisions, and establishes a disease evolution model that can reflect the development laws of the disease and the treatment effects. It uses the examinees' decisions and the current condition of the virtual patient as input to predict the development direction and degree of change of the disease. According to the prediction results of the disease evolution model, it adjusts the condition of the intelligent virtual patient in real time, and dynamically displays the development and changes of the disease by changing the symptoms and examination indicators of the virtual patient.

4. The clinical medicine practice skills assessment system according to claim 3, characterized in that: The disease dynamic control module establishes a disease evolution model that can reflect the disease development law and treatment effect. The model formula is: Among them, S t+1 is the predicted disease status value of the next time step, S t is the disease state value at the current time step, θ is the disease evolution coefficient, which controls the evolution speed of the disease with decision and time, and ΔS is the change in disease caused by the current decision. Its calculation formula is: ΔS is the change in the condition of the virtual patient after the candidate makes a decision, l is the number of decisions taken by the candidate, and a j is the effect coefficient of the jth decision, reflecting the impact of the decision on the condition, D j is the intensity factor of the jth decision, b j is the decision sensitivity coefficient, which controls the response speed of the disease to the decision, S0 is the current disease state value of the virtual patient, c j is the threshold of the j-th decision, is the random interference coefficient, which introduces randomness to simulate clinical uncertainty, ∈ is the random interference term, and obeys the probability distribution.

5. The clinical medicine practice skills assessment system according to claim 1, characterized in that: The assessment and analysis module constructs an evaluation index system covering multiple dimensions, including evaluation indicators of diagnostic accuracy, rationality of treatment plan, emergency response ability, communication ability, operational standardization, concentration and psychological state.

6. The clinical medicine practice skills assessment system according to claim 1, characterized in that: The assessment analysis module calculates the examinee's scores on each individual assessment indicator and the comprehensive score, and comprehensively evaluates the examinee's clinical practice skill level by adjusting and optimizing the indicator weights. The score calculation formula for each individual assessment indicator is: Among them, Score single is the score of the examinee on a single assessment indicator, p is the number of sub-assessment items under the single indicator, is the weight of the ith sub-evaluation item, which is determined according to the importance of each sub-evaluation item in the single indicator. is the candidate's matching score on the ith sub-assessment item, which is obtained by comparing the candidate's performance on this sub-assessment item with the standard. The comprehensive score formula is: Among them, Score total is the comprehensive score of the examinee, reflecting the examinee's clinical practice skill level, q is the total number of evaluation indicators, ω k is the weight of the k-th evaluation metric, is the examinee’s score on the kth evaluation indicator.

Images