Clinical thinking training method based on standardized cases

Abstract

The invention particularly relates to a clinical thinking training method based on standardized cases. According to the clinical thinking training method based on the standardized case, massive real medical data are collected, a medical knowledge graph is constructed by using natural language processing and machine learning technologies, and the standardized case is generated based on the medical knowledge graph; a clinical thinking training system is designed, and feedback and guidance are provided according to operation and input of students; constructing a clinical thinking evaluation model, performing quantitative evaluation on the operation data and the operation result of the trainee, generating an evaluation report, and analyzing a learning track; and customizing a personalized training path for the trainee by using a deep reinforcement learning algorithm, and performing dynamic adjustment in training. According to the clinical thinking training method based on the standardized cases, efficient and personalized clinical thinking training is provided for medical staff by constructing the standardized case library and combining an intelligent training system and an accurate evaluation feedback mechanism, and the clinical thinking ability and the practice level of the medical staff can be rapidly improved.

Description

Technical Field

[0001] The present invention relates to the field of medical teaching technology, and in particular to a clinical thinking training method based on standardized cases. Background Art

[0002] In medical education and clinical training, case construction technology is crucial for skills development, assessment, and clinical diagnostic support. Traditional standardized case construction methods rely primarily on long-term manual compilation and the experience of medical expert teams. However, due to the limitations of manual operation, this approach has numerous drawbacks and a high rate of data incompleteness. Furthermore, due to the limited number of experts and their limited time and energy, the case collection cycle is lengthy, making it difficult to meet the urgent needs of the ever-expanding medical education, the increasingly widespread medical applications, and the continuous deepening of scientific research.

[0003] As medical education moves toward digitalization, digital patient technology is gradually emerging. However, existing digital patient systems still have significant shortcomings in case construction and clinical thinking training.

[0004] During the case construction process, there is a lack of unified standardized specifications, and the generated cases often fail to truly and comprehensively reflect the complex and changing clinical reality.

[0005] In terms of clinical thinking training scenarios, the system only provides training for common case scenarios. There are insufficient training scenarios for rare diseases and complex and difficult cases, making it difficult to effectively cultivate students' ability to deal with complex and difficult cases.

[0006] In terms of training feedback mechanism, the system can only provide simple right and wrong judgments, and is unable to deeply analyze the causes of students' mistakes and provide targeted guidance and suggestions, resulting in a high repetition rate of the same mistakes in subsequent training.

[0007] Based on the above problems, it is urgent to develop a new clinical thinking training method based on standardized cases, which is of great significance for improving the quality of medical education and clinical training effects. Based on this, the present invention proposes a clinical thinking training method based on standardized cases. Summary of the Invention

[0008] In order to make up for the deficiencies of the prior art, the present invention provides a simple and efficient clinical thinking training method based on standardized cases.

[0009] The present invention is achieved through the following technical solutions:

[0010] A clinical thinking training method based on standardized cases, characterized by comprising the following steps:

[0011] Step S1: Constructing standardized cases

[0012] Collect massive amounts of real medical data, clean, annotate, and desensitize it, and then use natural language processing and machine learning technologies to build a medical knowledge graph. Based on the medical knowledge graph, standardized cases are generated, and after review by medical experts, a standardized case library is formed.

[0013] In step S1, the implementation process is as follows:

[0014] Step S1.1: Data collection

[0015] Collect patient data from hospital information systems and public datasets, medical literature, and clinical guidelines;

[0016] Step S1.2: Data preprocessing

[0017] Cleaning: Use data cleaning algorithms to remove duplicate, erroneous, and incomplete data records;

[0018] Desensitization: We adhere to strict data desensitization regulations and use encryption algorithms and hash functions to process sensitive patient information, including name, ID number, and contact information, to protect patient privacy and ensure that the data meets safety standards.

[0019] Annotation: Using professional annotation tools and based on unified medical annotation standards, we annotate medical entities in the data, including diseases, symptoms, and examination items, to provide a foundation for subsequent data analysis and model training.

[0020] Structuring: Using named entity recognition and relationship extraction methods in natural language processing technology, unstructured data is structured, key information is extracted, and customized conversion is performed into a format that is easy to store and process, laying the foundation for subsequent knowledge graph construction and case generation;

[0021] Step S1.3: Build a knowledge graph

[0022] Medical entity extraction: Using natural language processing and machine learning techniques, medical entities are identified from pre-processed data, including disease names, symptoms, examination items, and treatment plans.

[0023] Medical relationship extraction: Extracting causal and correlation relationships between medical entities through machine learning algorithms;

[0024] Constructing a knowledge graph: Based on the extracted medical entities and their relationships, a medical knowledge graph is constructed to provide strong knowledge support for case construction;

[0025] Establish a knowledge graph update mechanism: collect the latest medical research results, clinical guidelines, and case data in real time, and dynamically update the knowledge graph to ensure the timeliness and accuracy of knowledge;

[0026] Review the knowledge graph: Collaborate with medical experts to review and verify the knowledge graph to improve its reliability; Step S1.4, Generate Cases and Evaluation

[0027] Based on the medical knowledge graph, the patient's specific information is matched and inferred with the knowledge in the medical knowledge graph, and standardized cases are customized and generated. Medical experts are invited to review the generated standardized cases, and the standardized cases are optimized based on expert feedback to ensure case quality.

[0028] Step S2: Design a clinical thinking training system

[0029] Guided by the needs of medical education and clinical practice, customized training modules covering interviews, physical examinations, auxiliary examinations, diagnosis, and treatment are designed. A standardized case library is used to simulate real clinical scenarios, guiding trainees in case analysis and decision-making.

[0030] Using deep neural network models from speech recognition technology and semantic understanding algorithms from natural language processing technology, we custom-develop an intelligent interactive system that supports multiple interaction methods, including text input, voice interaction, and graphical user interface interaction, to adapt to the usage habits of different students and provide them with real-time feedback and guidance.

[0031] Based on the students' operations and inputs, the system provides feedback and guidance based on preset diagnosis and treatment rules and knowledge models, through data analysis and reasoning, to help students correct errors and optimize their ideas.

[0032] In step S2, the implementation process is as follows:

[0033] Step S2.1: Design training module

[0034] Customize the design of interview training modules, physical examination training modules, auxiliary examination training modules, diagnosis training modules, and treatment training modules based on the processes of medical education and clinical practice; each training module can be customized with several different case scenarios based on the type of department and disease;

[0035] Step S2.2: Realize intelligent interaction

[0036] Using speech recognition technology and natural language processing technology to realize intelligent interactive system; students interact with the system through voice or text, and the system accurately understands the students' intentions through intelligent interactive system and provides corresponding feedback and guidance suggestions;

[0037] Step S2.3: Intelligent feedback guidance

[0038] During the trainee's operation, the system analyzes the trainee's input and operation behavior in real time, and provides feedback and guidance based on the preset diagnosis and treatment rules and knowledge models;

[0039] If the student selects the wrong disease or treatment during diagnosis, the system will issue an error prompt.

[0040] Step S3: Establish an evaluation feedback mechanism

[0041] Combine clinical practice guidelines and expert experience to determine evaluation indicators and corresponding weights, use machine learning algorithms to build a clinical thinking evaluation model, conduct quantitative evaluation of trainees' operational data and results during training, generate evaluation reports and analyze learning trajectories;

[0042] In step S3, the evaluation indicators include the logical accuracy and coverage of keywords based on clinical thinking, the accuracy of diagnosis and the rationality score of treatment plan;

[0043] After the trainees complete the training, the system automatically generates a detailed evaluation report, pointing out the trainees’ strengths and weaknesses, and providing targeted improvement suggestions;

[0044] At the same time, the system records the students' learning trajectory, analyzes their learning habits and progress, and provides data support for subsequent personalized training.

[0045] In step S3, a decision tree algorithm in a machine learning algorithm is used to construct a clinical thinking evaluation model;

[0046] After the trainees complete the training, the system automatically collects the trainees' operational data, inputs it into the clinical thinking assessment model for analysis, and generates an assessment report; the report lists in detail the trainees' scores on each assessment indicator, points out the trainees' strengths and weaknesses, and provides suggestions for improvement.

[0047] Step S4: Develop a personalized training plan

[0048] Based on multimodal data collection and analysis, machine learning and artificial intelligence algorithms are used to evaluate the students' situation. Combined with the standardized case library and learning objectives, deep reinforcement learning algorithms are used to customize personalized training paths for students, and dynamic adjustments are made during training.

[0049] In step S4, the implementation process is as follows:

[0050] Step S4.1: Multimodal data acquisition

[0051] Multimodal data collection includes collecting students' operation behaviors, voice interactions, answering situations, basic information and learning background data;

[0052] Step S4.2: Accurate evaluation

[0053] Analyze multimodal data using machine learning and artificial intelligence algorithms, with evaluation indicators including learners' learning behavior, knowledge mastery, and skill performance;

[0054] Step S4.3: Personalized training path generation

[0055] Based on the evaluation results, a deep reinforcement learning algorithm is used to generate personalized training paths for students;

[0056] For students whose answer accuracy exceeds the custom threshold but whose practical operation time in their learning background is less than the custom threshold, additional practical operation training will be provided;

[0057] For students whose visual content exceeds a custom threshold when selecting learning materials, we will prioritize providing image and video learning resources.

[0058] Step S4.4: Dynamic Adjustment

[0059] During the training process, the trainees' performance is monitored in real time, and the training content and difficulty are dynamically adjusted according to the trainees' learning progress and mastery;

[0060] If a student's score exceeds a custom threshold in a case training of a certain difficulty level, the system will automatically increase the difficulty of the training;

[0061] If a student scores below the custom threshold for N consecutive times in case training of a certain difficulty, the system will automatically reduce the training difficulty and provide more learning materials and guidance suggestions.

[0062] A clinical thinking training system based on standardized cases, used to implement the above method, comprises:

[0063] The standardized case construction module is responsible for collecting massive amounts of real medical data. After cleaning, labeling, and desensitization, it uses natural language processing and machine learning technologies to construct a medical knowledge graph. Based on the medical knowledge graph, standardized cases are generated, and after review by medical experts, a standardized case database is formed.

[0064] The clinical thinking training system, which includes a medical interview training module, a physical examination training module, an auxiliary examination training module, a diagnosis training module, and a treatment training module, is responsible for simulating real clinical scenarios using a standardized case library, guided by the needs of medical education and clinical practice, to guide students in case analysis and decision-making. Based on the students' operations and input, and based on preset diagnosis and treatment rules and knowledge models, the system provides feedback and guidance through data analysis and reasoning, helping students correct errors and optimize their thinking.

[0065] The intelligent interactive system is developed using deep neural network models from speech recognition technology and semantic understanding algorithms from natural language processing technology. It supports text input, voice interaction, and graphical user interface interaction to adapt to the usage habits of different students and provide real-time feedback and guidance to students.

[0066] The evaluation and feedback module is responsible for determining evaluation indicators and corresponding weights based on clinical practice guidelines and expert experience, using machine learning algorithms to build a clinical thinking evaluation model, quantitatively evaluating the trainees' operational data and results during training, generating evaluation reports, and analyzing learning trajectories;

[0067] The personalized training module is responsible for collecting and analyzing multimodal data, using machine learning and artificial intelligence algorithms to evaluate the students' situation, combining standardized case libraries and learning goals, and using deep reinforcement learning algorithms to customize personalized training paths for students, and dynamically adjust them during training.

[0068] A clinical thinking training device based on standardized cases, characterized by comprising a memory and a processor; the memory is used to store a computer program, and the processor is used to implement the above-mentioned method steps when executing the computer program.

[0069] A readable storage medium, characterized in that: a computer program is stored on the readable storage medium, and the computer program implements the above method steps when executed by a processor.

[0070] The beneficial effects of the present invention are: the clinical thinking training method based on standardized cases can provide medical staff with efficient and personalized clinical thinking training, enhance the accuracy and objectivity of evaluation, help to quickly improve the clinical thinking ability and practical level of medical staff, and thus promote the modernization of medical education and clinical training. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0072] Attachment Figure 1 Schematic diagram of the clinical thinking training method based on standardized cases of the present invention. DETAILED DESCRIPTION

[0073] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present invention.

[0074] The clinical thinking training method based on standardized cases includes the following steps:

[0075] Step S1: Constructing standardized cases

[0076] Collect massive amounts of real medical data, clean, annotate, and desensitize it, and then use natural language processing and machine learning technologies to build a medical knowledge graph. Based on the medical knowledge graph, standardized cases are generated, and after review by medical experts, a standardized case library is formed.

[0077] In step S1, the implementation process is as follows:

[0078] Step S1.1: Data collection

[0079] Collect patient data from hospital information systems and public datasets, medical literature, and clinical guidelines;

[0080] Step S1.2: Data preprocessing

[0081] Cleaning: Use data cleaning algorithms to remove duplicate, erroneous, and incomplete data records. For example, identify and delete duplicate data by comparing key information such as patient ID and visit time. Correct erroneous data using data validation rules and logical verification algorithms. Supplement or mark missing data based on data characteristics and statistical laws.

[0082] Desensitization: We adhere to strict data desensitization regulations and use encryption algorithms and hash functions to process sensitive patient information, including name, ID number, and contact information, to protect patient privacy and ensure that the data meets safety standards.

[0083] Annotation: Using professional annotation tools and based on unified medical annotation standards, we annotate medical entities in the data, including diseases, symptoms, and examination items, to provide a foundation for subsequent data analysis and model training.

[0084] Structuring: Using named entity recognition and relationship extraction methods in natural language processing technology, unstructured data is structured, key information is extracted, and customized conversion is performed into a format that is easy to store and process, laying the foundation for subsequent knowledge graph construction and case generation;

[0085] Step S1.3: Build a knowledge graph

[0086] Medical entity extraction: Using natural language processing and machine learning techniques, medical entities are identified from pre-processed data, including disease names, symptoms, examination items, and treatment plans.

[0087] Medical relationship extraction: Extracting causal and correlation relationships between medical entities through machine learning algorithms; for example, the relationship between diseases and symptoms, the correspondence between diseases and treatment plans, etc.

[0088] For example, by analyzing medical record texts, the relationship between diseases and symptoms can be determined, such as the relationship between "cold" and symptoms such as "fever, cough, runny nose".

[0089] Constructing a knowledge graph: Based on the extracted medical entities and their relationships, a medical knowledge graph is constructed to provide strong knowledge support for case construction;

[0090] Establish a knowledge graph update mechanism: Collect the latest medical research results, clinical guidelines, and case data in real time, and dynamically update the knowledge graph. By continuously updating and improving the medical knowledge graph, incorporating the latest medical research results and clinical practice experience, we ensure the timeliness and accuracy of knowledge.

[0091] Review the knowledge graph: Collaborate with medical experts to review and verify the knowledge graph to improve its reliability; Step S1.4, Generate Cases and Evaluation

[0092] Based on the medical knowledge graph, the patient's specific information is matched and inferred with the knowledge in the medical knowledge graph, and standardized cases are customized and generated. Medical experts are invited to review the generated standardized cases, and the standardized cases are optimized based on expert feedback to ensure case quality.

[0093] Step S2: Design a clinical thinking training system

[0094] Guided by the needs of medical education and clinical practice, customized training modules covering interviews, physical examinations, auxiliary examinations, diagnosis, and treatment are designed. A standardized case library is used to simulate real clinical scenarios, guiding trainees in case analysis and decision-making.

[0095] Using deep neural network models from speech recognition technology and semantic understanding algorithms from natural language processing technology, we custom-develop an intelligent interactive system that supports multiple interaction methods, including text input, voice interaction, and graphical user interface interaction, to adapt to the usage habits of different students and provide them with real-time feedback and guidance.

[0096] Based on the students' operations and inputs, the system provides feedback and guidance based on preset diagnosis and treatment rules and knowledge models, through data analysis and reasoning, to help students correct errors and optimize their ideas.

[0097] For example, when students ask inappropriate questions during the consultation process, the system will promptly prompt the correct direction of consultation based on semantic understanding and knowledge base matching; in the diagnosis process, if the student's diagnosis results are inaccurate, the system will analyze the reasons and provide relevant diagnostic basis and guidance.

[0098] In step S2, the implementation process is as follows:

[0099] Step S2.1: Design training module

[0100] The interview training module, physical examination training module, auxiliary examination training module, diagnosis training module and treatment training module are customized according to the process of medical education and clinical practice; each training module is customized with several different types of case scenarios according to the type of department and disease, such as heart disease and diabetes cases in internal medicine, fractures and appendicitis cases in surgery, etc.

[0101] Step S2.2: Realize intelligent interaction

[0102] Speech recognition technology and natural language processing technology are used to realize an intelligent interactive system; students interact with the system through voice or text, and the system accurately understands the students' intentions through the intelligent interactive system and provides corresponding feedback and guidance suggestions; for example, if the student says "the patient has cough symptoms", the system automatically displays cough-related medical questions or examination suggestions.

[0103] Step S2.3: Intelligent feedback guidance

[0104] During the trainee's operation, the system analyzes the trainee's input and operation behavior in real time, and provides feedback and guidance based on the preset diagnosis and treatment rules and knowledge models;

[0105] If the student selects the wrong disease or treatment during diagnosis, the system will issue an error message: "Based on the current symptoms and test results, the diagnosis is not accurate. Please re-analyze and consider other possible diseases," and provide relevant diagnostic ideas.

[0106] Step S3: Establish an evaluation feedback mechanism

[0107] Combine clinical practice guidelines and expert experience to determine evaluation indicators and corresponding weights, use machine learning algorithms to build a clinical thinking evaluation model, conduct quantitative evaluation of trainees' operational data and results during training, generate evaluation reports and analyze learning trajectories;

[0108] In step S3, the evaluation indicators include the logical accuracy and coverage of keywords based on clinical thinking, the accuracy of diagnosis and the rationality score of the treatment plan (the system scores the rationality of the treatment plan according to the built-in custom scoring criteria);

[0109] For example, the weight of the diagnostic accuracy is set to 0.4, the weight of the rationality score of the treatment plan is set to 0.3, and the weight of the logical accuracy and coverage of keywords based on clinical thinking is set to 0.3.

[0110] After the trainee completes the training, the system automatically generates a detailed evaluation report, pointing out the trainee's strengths (items where the evaluation indicators exceed the custom threshold) and weaknesses (items where the evaluation indicators are below the custom threshold), and provides targeted improvement suggestions;

[0111] For example, if a trainee overlooks certain important symptoms during the diagnosis process, the evaluation report will clearly point this out and suggest that the trainee strengthen their ability to analyze the correlation between symptoms.

[0112] At the same time, the system records the students' learning trajectory, analyzes their learning habits and progress, and provides data support for subsequent personalized training.

[0113] For example, if it is found that students often make errors in diagnosing a certain type of disease, the system will add case training for this type of disease in subsequent training.

[0114] In step S3, a decision tree algorithm in a machine learning algorithm is used to construct a clinical thinking evaluation model;

[0115] After the trainees complete the training, the system automatically collects the trainees' operational data, inputs it into the clinical thinking assessment model for analysis, and generates an assessment report; the report lists in detail the trainees' scores on each assessment indicator, points out the trainees' strengths and weaknesses, and provides suggestions for improvement.

[0116] For example, if a student overlooks certain important symptoms during the diagnosis process, the assessment report will clearly point this out and recommend that the student strengthen their ability to analyze symptom correlations. The system also records the student's learning trajectory, analyzes their learning habits and progress, and provides data support for subsequent personalized training. For example, if a student is found to frequently make errors in diagnosing a certain disease, the system will add case studies of this disease to subsequent training.

[0117] Step S4: Develop a personalized training plan

[0118] Based on multimodal data collection and analysis, machine learning and artificial intelligence algorithms are used to evaluate the students' situation. Combined with the standardized case library and learning objectives, deep reinforcement learning algorithms are used to customize personalized training paths for students, and dynamic adjustments are made during training.

[0119] In step S4, the implementation process is as follows:

[0120] Step S4.1: Multimodal data acquisition

[0121] Multimodal data collection includes collecting students' operation behaviors, voice interactions, answering situations, basic information and learning background data;

[0122] Step S4.2: Accurate evaluation

[0123] Analyze multimodal data using machine learning and artificial intelligence algorithms, with evaluation indicators including learners' learning behavior, knowledge mastery, and skill performance;

[0124] For example, by analyzing the students' answer accuracy and answering time, we can evaluate their mastery of different knowledge points.

[0125] Step S4.3: Personalized training path generation

[0126] Based on the evaluation results, a deep reinforcement learning algorithm is used to generate personalized training paths for students;

[0127] For students whose answer accuracy exceeds the custom threshold but whose practical operation time in their learning background is less than the custom threshold, additional practical operation training will be provided;

[0128] For students whose visual content exceeds a custom threshold when selecting learning materials, we will prioritize providing image and video learning resources.

[0129] When customizing personalized training paths, the system prioritizes simple case training for students with weak foundations, gradually increasing the difficulty. For students with a solid foundation in a specific field (such as internal medicine), training on more complex cases in that field is provided. During training, the system dynamically adjusts the content and difficulty based on the student's real-time performance to ensure the consistency and effectiveness of the learning path. The system integrates multimodal learning resources, such as text, images, and videos, to meet the needs of students with different learning styles, forming a continuously optimized learning loop.

[0130] Step S4.4: Dynamic Adjustment

[0131] During the training process, the trainees' performance is monitored in real time, and the training content and difficulty are dynamically adjusted according to the trainees' learning progress and mastery;

[0132] If a student's score exceeds a custom threshold in a case training of a certain difficulty level, the system will automatically increase the difficulty of the training;

[0133] If a student scores below the custom threshold for N consecutive times in case training of a certain difficulty, the system will automatically reduce the training difficulty and provide more learning materials and guidance suggestions.

[0134] The clinical thinking training system based on standardized cases is used to implement the above method, including:

[0135] The standardized case construction module is responsible for collecting massive amounts of real medical data. After cleaning, labeling, and desensitization, it uses natural language processing and machine learning technologies to construct a medical knowledge graph. Based on the medical knowledge graph, standardized cases are generated, and after review by medical experts, a standardized case database is formed.

[0136] The clinical thinking training system, which includes a medical interview training module, a physical examination training module, an auxiliary examination training module, a diagnosis training module, and a treatment training module, is responsible for simulating real clinical scenarios using a standardized case library, guided by the needs of medical education and clinical practice, to guide students in case analysis and decision-making. Based on the students' operations and input, and based on preset diagnosis and treatment rules and knowledge models, the system provides feedback and guidance through data analysis and reasoning, helping students correct errors and optimize their thinking.

[0137] The intelligent interactive system is developed using deep neural network models from speech recognition technology and semantic understanding algorithms from natural language processing technology. It supports text input, voice interaction, and graphical user interface interaction to adapt to the usage habits of different students and provide real-time feedback and guidance to students.

[0138] The evaluation and feedback module is responsible for determining evaluation indicators and corresponding weights based on clinical practice guidelines and expert experience, using machine learning algorithms to build a clinical thinking evaluation model, quantitatively evaluating the trainees' operational data and results during training, generating evaluation reports, and analyzing learning trajectories;

[0139] The personalized training module is responsible for collecting and analyzing multimodal data, using machine learning and artificial intelligence algorithms to evaluate the students' situation, combining standardized case libraries and learning goals, and using deep reinforcement learning algorithms to customize personalized training paths for students, and dynamically adjust them during training.

[0140] The clinical thinking training device based on standardized cases includes a memory and a processor; the memory is used to store a computer program, and the processor is used to implement the above method steps when executing the computer program.

[0141] The readable storage medium stores a computer program, which implements the above method steps when executed by a processor.

[0142] Compared with existing technologies, this clinical thinking training method based on standardized cases has the following characteristics:

[0143] (1) Improved training efficiency and quality: The standardized case library provides rich, reusable training resources, allowing trainees to train in a variety of clinical scenarios, improving their familiarity with and ability to respond to various cases. The intelligent interactive system and real-time feedback mechanism enable trainees to receive timely guidance, avoid the accumulation of errors, and accelerate the improvement of clinical thinking ability.

[0144] (2) Personalized training is achieved: Through accurate assessment of students and the development of personalized training plans, the learning needs of different students can be met and learning outcomes can be improved. Whether students have a weak foundation or have some experience, they can train at a difficulty and pace that suits them and give full play to their learning potential.

[0145] (3) Enhanced accuracy and objectivity of evaluation: Quantified evaluation indicators and automatically generated evaluation reports make the evaluation process more scientific and objective, and can accurately reflect the students' clinical thinking level and ability. This helps students understand their own learning status and provides teachers with an objective basis for teaching evaluation, facilitating the adjustment of teaching strategies.

[0146] (4) Promoting the development of medical education and clinical training: The method of the present invention provides an innovative solution for medical education and clinical training, which helps to cultivate more high-quality medical staff, improve the quality and safety of medical services, and promote the modernization of medical education and clinical training.

[0147] The embodiment described above is only one specific implementation of the present invention. Common changes and substitutions made by those skilled in the art within the scope of the technical solution of the present invention should be included in the protection scope of the present invention.

Claims

1. A clinical thinking training method based on standardized cases, characterized by: The following steps are involved: Step S1: Constructing standardized cases Collect massive amounts of real medical data, clean, annotate, and desensitize it, and then use natural language processing and machine learning technologies to build a medical knowledge graph. Based on the medical knowledge graph, standardized cases are generated, and after review by medical experts, a standardized case database is formed. Step S2: Design a clinical thinking training system Guided by the needs of medical education and clinical practice, customized training modules covering interviews, physical examinations, auxiliary examinations, diagnosis, and treatment are designed. A standardized case library is used to simulate real clinical scenarios, guiding trainees in case analysis and decision-making. Using deep neural network models from speech recognition technology and semantic understanding algorithms from natural language processing technology, we custom-develop an intelligent interactive system that supports text input, voice interaction, and a graphical user interface to adapt to the usage habits of different students and provide real-time feedback and guidance to them. Based on the students' operations and inputs, the system provides feedback and guidance by implementing data analysis and reasoning based on preset diagnosis and treatment rules and knowledge models, helping students correct errors and optimize their thinking. Step S3: Establish an evaluation feedback mechanism Combine clinical practice guidelines and expert experience to determine evaluation indicators and corresponding weights, use machine learning algorithms to build a clinical thinking evaluation model, conduct quantitative evaluation of trainees' operational data and results during training, generate evaluation reports and analyze learning trajectories; Step S4: Develop a personalized training plan Based on multimodal data collection and analysis, machine learning and artificial intelligence algorithms are used to evaluate the students' situation. Combined with the standardized case library and learning objectives, deep reinforcement learning algorithms are used to customize personalized training paths for students, and dynamic adjustments are made during training.

2. The clinical thinking training method based on standardized cases according to claim 1, characterized in that: In step S1, the implementation process is as follows: Step S1.1: Data collection Collect patient data from hospital information systems and public datasets, medical literature, and clinical guidelines; Step S1.2: Data preprocessing Cleaning: Use data cleaning algorithms to remove duplicate, erroneous, and incomplete data records; Desensitization: We follow data desensitization standards and use encryption algorithms and hash functions to process patient sensitive information, including name, ID number, and contact information, to protect patient privacy and ensure that the data meets safety standards. Annotation: Using annotation tools and based on unified medical annotation standards, medical entities in the data, including diseases, symptoms, and examination items, are annotated to provide a basis for subsequent data analysis and model training. Structuring: Using named entity recognition and relationship extraction methods in natural language processing technology, unstructured data is structured, key information is extracted, and customized data formats are converted to lay the foundation for subsequent knowledge graph construction and case generation; Step S1.3: Build a knowledge graph Medical entity extraction: Using natural language processing and machine learning techniques, medical entities are identified from pre-processed data, including disease names, symptoms, examination items, and treatment plans. Medical relationship extraction: Extracting causal and correlation relationships between medical entities through machine learning algorithms; Constructing a knowledge graph: Based on the extracted medical entities and their relationships, construct a medical knowledge graph; Establish a knowledge graph update mechanism: collect the latest medical research results, clinical guidelines, and case data in real time, and dynamically update the knowledge graph to ensure the timeliness and accuracy of knowledge; Review the knowledge graph: By collaborating with medical experts, the knowledge graph will be reviewed and verified to improve its reliability; Step S1.4: Generate cases and evaluate Based on the medical knowledge graph, the patient's specific information is matched and inferred with the knowledge in the medical knowledge graph, and standardized cases are customized and generated. Medical experts are invited to review the generated standardized cases, and the standardized cases are optimized based on expert feedback to ensure case quality.

3. The clinical thinking training method based on standardized cases according to claim 1, characterized in that: In step S2, the implementation process is as follows: Step S2.1: Design training module Customize the design of interview training modules, physical examination training modules, auxiliary examination training modules, diagnosis training modules, and treatment training modules based on the processes of medical education and clinical practice; each training module can be customized with several different types of case scenarios according to the type of department and disease; Step S2.2: Realize intelligent interaction Using speech recognition technology and natural language processing technology to realize intelligent interactive system; students interact with the system through voice or text, and the system accurately understands the students' intentions through the intelligent interactive system; Step S2.3: Intelligent feedback guidance During the trainee's operation, the system analyzes the trainee's input and operation behavior in real time, and provides feedback and guidance based on the preset diagnosis and treatment rules and knowledge models; If the student selects the wrong disease or treatment during diagnosis, the system will issue an error prompt.

4. The clinical thinking training method based on standardized cases according to claim 1, characterized in that: In step S3, the evaluation indicators include the logical accuracy and coverage of keywords based on clinical thinking, the accuracy of diagnosis and the rationality score of treatment plan; After the trainees complete the training, the system automatically generates a detailed evaluation report, pointing out the trainees’ strengths and weaknesses, and providing targeted improvement suggestions; At the same time, the system records the students' learning trajectory, analyzes their learning habits and progress, and provides data support for subsequent personalized training.

5. The clinical thinking training method based on standardized cases according to claim 4, characterized in that: In step S3, a decision tree algorithm in a machine learning algorithm is used to construct a clinical thinking evaluation model; After the trainees complete the training, the system automatically collects the trainees' operational data, inputs it into the clinical thinking assessment model for analysis, and generates an assessment report; the report lists in detail the trainees' scores on each assessment indicator, points out the trainees' strengths and weaknesses, and provides suggestions for improvement.

6. The clinical thinking training method based on standardized cases according to claim 1, characterized in that: In step S4, the implementation process is as follows: Step S4.1: Multimodal data acquisition Multimodal data collection includes collecting students' operation behaviors, voice interactions, answering situations, basic information and learning background data; Step S4.2: Accurate evaluation Analyze multimodal data using machine learning and artificial intelligence algorithms, with evaluation indicators including learners' learning behavior, knowledge mastery, and skill performance; Step S4.3: Generate personalized training path Based on the evaluation results, a deep reinforcement learning algorithm is used to generate a personalized training path for the students; For students whose answer accuracy exceeds the custom threshold but whose practical operation time in their learning background is less than the custom threshold, additional practical operation training will be provided; For students whose visual content exceeds a custom threshold when selecting learning materials, we will prioritize providing image and video learning resources. Step S4.4: Dynamic Adjustment During the training process, the trainees' performance is monitored in real time, and the training content and difficulty are dynamically adjusted according to the trainees' learning progress and mastery; If a student's score exceeds a custom threshold in a case training of a certain difficulty level, the system will automatically increase the difficulty of the training; If a student scores below the custom threshold for N consecutive times in case training of a certain difficulty, the system will automatically reduce the training difficulty and provide more learning materials and guidance suggestions.

7. A clinical thinking training system based on standardized cases, characterized by: The method for implementing any one of claims 1 to 6 comprises: The standardized case construction module is responsible for collecting massive amounts of real medical data. After cleaning, labeling, and desensitization, it uses natural language processing and machine learning technologies to construct a medical knowledge graph. Based on the medical knowledge graph, standardized cases are generated, and after review by medical experts, a standardized case database is formed. The clinical thinking training system, which includes a medical interview training module, a physical examination training module, an auxiliary examination training module, a diagnosis training module, and a treatment training module, is responsible for simulating real clinical scenarios using a standardized case library, guided by the needs of medical education and clinical practice, to guide students in case analysis and decision-making. Based on the students' operations and input, and based on preset diagnosis and treatment rules and knowledge models, the system provides feedback and guidance through data analysis and reasoning, helping students correct errors and optimize their thinking. The intelligent interactive system is developed using deep neural network models from speech recognition technology and semantic understanding algorithms from natural language processing technology. It supports text input, voice interaction, and graphical user interface interaction to adapt to the usage habits of different students and provide real-time feedback and guidance to students. The evaluation and feedback module is responsible for determining evaluation indicators and corresponding weights based on clinical practice guidelines and expert experience, using machine learning algorithms to build a clinical thinking evaluation model, quantitatively evaluating the trainees' operational data and results during training, generating evaluation reports, and analyzing learning trajectories; The personalized training module is responsible for collecting and analyzing multimodal data, using machine learning and artificial intelligence algorithms to evaluate the students' situation, combining standardized case libraries and learning goals, and using deep reinforcement learning algorithms to customize personalized training paths for students, and dynamically adjust them during training.

8. A clinical thinking training device based on standardized cases, characterized by: The method comprises a memory and a processor; the memory is used to store a computer program, and the processor is used to implement the method steps according to any one of claims 1 to 6 when executing the computer program.

9. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which, when executed by a processor, implements the method steps according to any one of claims 1 to 6.

Images