A large language-based oral medical student reception and medical record writing teaching system and method

Abstract

This invention relates to the field of medical education technology and discloses a teaching system and method for oral medicine students' patient consultation and medical record writing based on a large language model. The system includes a case database module, an AI virtual patient interaction module, an AI scoring feedback module, and a learning data recording and analysis module. This teaching system for oral medicine students' patient consultation and medical record writing, based on a large language model, establishes unified and objective multi-dimensional scoring standards to accurately score each item of the resident physician's consultation dialogues and medical record writing content. This effectively eliminates the inconsistency in standards caused by subjective judgment differences in traditional manual evaluation. The system can generate a detailed scoring report immediately after each practice session, clearly pointing out existing problems and deductions, and providing specific improvement suggestions and standardized demonstrations for each problem. This enables resident physicians to correct errors promptly and in a targeted manner, shortening the learning feedback cycle.

Description

Technical Field

[0001] This invention relates to the field of medical education technology, and in particular to a teaching system and method for oral medicine students to receive patients and write medical records based on a large language. Background Technology

[0002] Clinical patient reception skills and medical record writing proficiency are core training objectives for dental students. Traditional training for dental residents in patient reception and medical record writing primarily relies on theoretical learning, clinical demonstrations by supervising physicians, and limited patient practice. However, with the increasing demand for medical education and the scarcity of clinical teaching resources, the limitations of the traditional training model are becoming increasingly apparent.

[0003] 1. Limited training resources and incomplete case coverage.

[0004] Traditional training relies heavily on paper-based case databases or limited electronic case systems. The types and number of cases are constrained by the actual scope of clinical practice. While there are numerous subspecialties and complex diseases in dentistry, there are relatively ample opportunities for practice with common and frequently occurring diseases. However, training materials for rare and special cases are severely lacking, resulting in residents' insufficient experience when facing complex or atypical cases. Furthermore, dentistry knowledge updates rapidly, with new technologies and guidelines constantly emerging, while traditional case data lags significantly in updating, making it difficult to keep pace with the forefront of clinical practice.

[0005] 2. Insufficient ability to simulate real-world scenarios

[0006] Traditional training often relies on static written case analyses or simple scenario simulations, lacking real-time, dynamic interaction with "patients." In real clinical settings, patients' expressions, symptom descriptions, and emotional responses are highly unpredictable, requiring physicians to flexibly adjust their interviewing strategies and diagnostic approaches. Traditional models cannot simulate this dynamic and unpredictable nature of interaction, leading to communication difficulties for resident physicians when facing real patients alone. Furthermore, the real clinical environment presents complex interfering factors such as equipment noise, patient anxiety, and time pressure, which traditional training environments struggle to replicate, hindering the development of resident physicians' resilience and adaptability.

[0007] 3. The evaluation and feedback mechanism is inadequate.

[0008] Traditional teaching evaluation primarily relies on manual assessment by supervising physicians. Due to differences in professional background, clinical experience, and subjective preferences among instructors, evaluation standards are difficult to standardize and quantify: some instructors emphasize the completeness of the medical history, while others prioritize the standardization of medical record formats, making it difficult for resident physicians to obtain consistent and clear directions for improvement. Furthermore, the busy schedules of clinical supervising physicians often lead to time delays in feedback, causing resident physicians to miss optimal corrective opportunities. Manual assessment also struggles to provide detailed, item-by-item analysis, making it prone to oversights and failing to comprehensively and accurately pinpoint the specific problems and shortcomings of resident physicians.

[0009] 4. Lack of data management and analysis capabilities

[0010] Traditional training data is mostly stored in paper or scattered electronic files, which is difficult to collect and integrate, consumes a lot of manpower and time, and is prone to loss or incomplete information, making it difficult to support a systematic evaluation of training effectiveness. At the same time, the lack of advanced data analysis technology makes it impossible to deeply explore the common weaknesses, individual learning progress and differentiated learning characteristics of resident physicians from massive amounts of training data, and the optimization of training programs lacks a data-driven scientific basis.

[0011] 5. Special needs of the dental profession were not met.

[0012] Oral medicine has unique professional skill requirements, such as the standardization of oral examination techniques and the comprehensive consideration of the relationship between local and systemic diseases. Traditional training often emphasizes theory over practice, failing to fully integrate specialized examination techniques, special diagnostic and treatment procedures with patient history taking and medical record writing. As a result, resident physicians struggle to develop comprehensive oral clinical thinking during practice, affecting their subsequent practical diagnostic and treatment abilities and the quality of their medical record writing. Summary of the Invention

[0013] In view of the problems of limited training resources, incomplete case coverage, insufficient ability to simulate real scenarios, imperfect evaluation and feedback mechanisms, lack of data management and analysis capabilities, and failure to meet the special needs of the dental profession in the existing traditional training model, this invention is proposed.

[0014] Therefore, the purpose of this invention is to provide a teaching system for dental students to receive patients and write medical records based on a large language, which aims to provide a rich variety of standard patients, realize real-time interactive simulation, have objective and unified evaluation standards, conduct in-depth analysis of learning data, and fully consider the training needs of dental professional skills.

[0015] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a teaching system for dental students' patient reception and medical record writing based on big language, comprising:

[0016] The case database module stores standard case data for various oral diseases. The standard case data includes at least the patient's basic information, chief complaint, present illness history, systemic medical history, and oral examination results.

[0017] The AI ​​virtual patient interaction module, built on a large language model, is used to simulate a virtual standard patient based on the standard case data, generate responses that conform to the actual clinical situation in real time based on the consultation content input by the learner, and conduct simulated consultation dialogue with the learner.

[0018] The AI ​​scoring feedback module, built on a large language model and preset multi-dimensional scoring standards, is used to automatically score the consultation dialogue records and subsequent medical records generated by the learner during the consultation simulation, and output the scoring results and corresponding improvement suggestions.

[0019] The learning data recording and analysis module is used to automatically record the complete process data and multi-dimensional scoring results of each practice session of the learner, and analyze the data to generate a learning progress analysis report for the learner.

[0020] As a preferred embodiment of the teaching system for oral medicine students to receive patients and write medical records based on big language as described in this invention, the AI ​​virtual patient interaction module is configured to simulate standard virtual patients in the dental department of different ages, different chief complaints, different oral diseases and different emotional states, and dynamically adjust the response content according to the learner's consultation strategy.

[0021] As a preferred embodiment of the teaching system for oral medicine students' patient reception and medical record writing based on big language as described in this invention, the multi-dimensional scoring criteria used by the AI ​​scoring feedback module include: medical history collection scoring dimension, oral examination scoring dimension, medical record writing scoring dimension, diagnosis and treatment plan scoring dimension, and doctor-patient communication scoring dimension.

[0022] As a preferred embodiment of the teaching system for oral medicine students' patient reception and medical record writing based on big language as described in this invention, the medical record writing scoring dimensions are further subdivided into at least three sub-dimensions: completeness of chief complaint, logicality of present medical history, completeness of systemic medical history, standardization of oral examination description, accuracy of diagnosis, and rationality of treatment plan.

[0023] As a preferred embodiment of the teaching system for oral medicine students' patient reception and medical record writing based on big language as described in this invention, the AI ​​scoring feedback module is further configured to provide corresponding correct demonstration examples and improvement suggestions for deducted points while generating scoring results.

[0024] As a preferred embodiment of the teaching system for oral medicine students' consultation and medical record writing based on big language as described in this invention, the learning data recording and analysis module is further configured to: automatically identify the learner's weak knowledge areas and ability progress trends through longitudinal comparative analysis of the practice data, and generate a visual learning curve report.

[0025] As a preferred embodiment of the present invention, the step includes the following steps:

[0026] S1. Case push: Select or push standard cases of the target oral disease from the case database to the learner;

[0027] S2, AI-assisted patient reception simulation: The learner, based on the pushed standard case information, conducts real-time intelligent dialogue with a virtual standard patient through the AI ​​virtual patient interaction module to carry out simulated practice of consultation and medical history collection;

[0028] S3. Medical Record Writing: After the simulated consultation, the learner completes the medical record writing within the system based on the information obtained from the simulated consultation.

[0029] S4. Scoring and Feedback: The AI ​​scoring and feedback module automatically scores the learner's consultation dialogue records and medical records from multiple dimensions, and outputs a feedback report containing the scoring results and improvement suggestions in real time.

[0030] S5. Learning Data Recording and Analysis: The learning data recording and analysis module records all process data and scoring results of this exercise, and updates the learner's personal learning progress analysis report.

[0031] In a preferred embodiment of the present invention, in step S2, the AI ​​virtual patient interaction module generates a response by intelligently analyzing and matching the questions input by the learner based on the preset medical condition information in the standard case data, combined with the oral medicine professional knowledge base and clinical diagnosis and treatment logic.

[0032] As a preferred embodiment of the present invention, in step S4, the AI ​​scoring feedback module compares the learner's consultation dialogue record and medical records with the preset scoring points item by item, and gives scores and text comments for each item according to the professional standards of oral medicine.

[0033] Compared with the prior art, the present invention has at least the following beneficial effects:

[0034] 1. This invention establishes a unified and objective multi-dimensional scoring standard based on a large language model to accurately score each item of resident physicians' consultation dialogues and medical record writing. This effectively eliminates the problem of inconsistent standards caused by subjective judgment differences in traditional manual evaluation. The system can generate a detailed scoring report immediately after each practice session, which not only clearly points out the existing problems and deductions, but also provides specific improvement suggestions and standardized demonstrations for each problem, enabling resident physicians to correct errors in a timely and targeted manner and shorten the learning feedback cycle.

[0035] 2. This invention utilizes a large language model to construct a virtual standard patient capable of real-time intelligent dialogue. It can dynamically generate diverse and clinically relevant responses based on the resident physician's consultation content, simulating different patients' chief complaint expression habits, emotional reactions, and symptom changes. The system can cover common, frequently occurring, and rare oral diseases, effectively compensating for the shortcomings of traditional training in terms of limited case types and insufficient real-time interaction, and helping resident physicians develop flexible adaptability and clinical communication skills in a simulated environment.

[0036] 3. This invention can automatically record the complete process data of each resident physician's practice, including consultation dialogue records, medical record writing content, scores for each dimension and each item. Through systematic integration and in-depth analysis of learning data, it can automatically identify the common weaknesses of resident physicians and individual differences in learning progress, and generate personal learning progress reports, providing objective and quantitative data support for the formulation of personalized training programs and the adjustment of the overall teaching plan. Attached Figure Description

[0037] Figure 1 This is a schematic diagram illustrating the method steps of the teaching system for oral medicine students' patient reception and medical record writing based on big language, as described in this invention.

[0038] Figure 2 This is a schematic diagram of the system flow of the teaching system for oral medicine students' patient reception and medical record writing based on big language, which is based on the present invention. Detailed Implementation

[0039] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0040] Reference Figures 1-2 This invention provides a teaching system for dental students' patient reception and medical record writing based on a large language framework, comprising the following core modules:

[0041] (a) Case Database Module

[0042] This module stores standard case data for various oral diseases. The standard case data includes at least basic patient information, chief complaint, present illness, systemic medical history, and oral examination results. The case database covers common, frequently occurring, and some rare diseases across various subspecialties of dentistry, including but not limited to typical conditions such as pulpitis, periodontitis, and impacted teeth. The case data is stored in a structured format for easy system retrieval and AI model analysis. The case database supports dynamic updates, allowing for timely addition of new cases based on the latest dentistry guidelines and clinical advancements.

[0043] (ii) AI Virtual Patient Interaction Module

[0044] This module, built upon a large language model, is used to simulate virtual standard patients based on the standard case data. The large language model, fine-tuned with expertise in oral medicine, can generate responses in real-time that conform to actual clinical situations based on the learner's input of consultation questions, enabling simulated patient dialogue with the learner.

[0045] This module is configured to simulate virtual standard patients in dentistry of different ages, with different chief complaints, different oral diseases, and different emotional states. For example, in a pulpitis case, the virtual patient can simulate the clinical characteristics of acute pulpitis patients who are unable to express themselves clearly due to severe pain and are emotionally agitated; in an impacted tooth case, it can simulate the anxiety and fear of a young patient undergoing tooth extraction. The virtual patient can also dynamically adjust its responses based on the learner's questioning strategy—if the learner's questioning is too brief, the virtual patient will proactively add details of the symptoms or raise questions; if the learner uses overly technical terminology, the virtual patient may show incomprehension, prompting the learner to provide a simplified explanation, thereby improving the learner's clinical communication skills.

[0046] (III) AI Scoring Feedback Module

[0047] This module is built on a large language model and a preset multi-dimensional scoring standard. It is used to automatically score the consultation dialogue records and subsequent medical records generated by learners during the consultation simulation, and output the scoring results and corresponding improvement suggestions.

[0048] The multi-dimensional scoring criteria are formulated based on the professional standards and clinical teaching requirements of stomatology, and mainly include the following dimensions:

[0049] Medical history taking scoring dimensions: assessing the learner's comprehensiveness, systematicness, and logic in taking medical history;

[0050] Oral examination scoring dimensions: assessing the rationality and completeness of the selection of oral examination items;

[0051] The scoring dimensions for medical record writing are further subdivided into subdivisions such as completeness of the chief complaint, logicality of the present illness history, completeness of the systemic medical history, standardization of the description of oral examination, accuracy of diagnosis, and rationality of the treatment plan.

[0052] Diagnosis and Treatment Plan Scoring Dimensions: Assessing the correctness of the diagnostic logic and the scientific validity of the treatment plan;

[0053] The scoring dimensions for doctor-patient communication include: assessing communication skills, humanistic care, and the standardization and friendliness of language expression.

[0054] The AI ​​scoring feedback module is also configured to provide a corresponding correct example and improvement suggestion text for each deducted item while generating the scoring results, so that learners not only know "where the points were deducted", but also understand "why the points were deducted" and "how to improve".

[0055] (iv) Learning Data Recording and Analysis Module

[0056] This module automatically records complete process data and multi-dimensional scoring results for each learner's practice session. The process data includes the full text of the consultation dialogue, the original medical record, the consultation time, and scores and deduction details for each dimension. The module is also configured to automatically identify learners' weak knowledge areas and skill improvement trends through longitudinal comparative analysis of practice data, generating a visualized learning curve report. The analysis results can be used to develop personalized training plans for each learner and provide objective, quantitative data support for optimizing and adjusting the overall teaching plan.

[0057] Example 2

[0058] A teaching method for dental students in patient reception and medical record writing is also provided, which is applied to the aforementioned teaching system. This method includes the following steps:

[0059] S1, Case Push

[0060] The system selects or pushes standard cases of target oral diseases from the case database to learners. In one embodiment, the system may randomly assign cases; in another embodiment, the system may intelligently recommend specific disease types that require more practice based on the learner's historical practice data and weak disease areas, thereby achieving targeted training.

[0061] The pushed case information is presented in a standardized outpatient medical record format, including: basic information such as the patient's name, age, and gender; chief complaint (e.g., "spontaneous pain in the lower right posterior tooth for three days"); and a brief description of the background of the visit. Learners first review the case information to understand the patient's basic situation and the reason for the visit, preparing for subsequent consultation simulations.

[0062] S2, AI-assisted patient reception simulation

[0063] Based on the standard case information pushed in step S1, learners engage in real-time intelligent dialogue with virtual standard patients through the AI ​​virtual patient interaction module, and conduct simulated practice of consultation and medical history collection.

[0064] In this step, the learner, acting as a physician, inputs consultation questions through the system interface. When generating a response, the AI ​​virtual patient interaction module intelligently analyzes and matches the learner's input questions based on preset medical information from the standard case data, combined with a professional knowledge base of oral medicine and clinical diagnosis logic.

[0065] For example:

[0066] The learner asked, "How long have you had a toothache?"

[0067] The virtual patient replied: "It's been about three days. At first, it hurt when I ate, but these past two days, the pain has been intermittent even when I haven't eaten. Last night, the pain was so bad that I couldn't sleep well."

[0068] The learner asked, "Could you point out which tooth hurts?"

[0069] The virtual patient replied, "It feels like it's a tooth on the lower right side, towards the back, but I'm not quite sure which one." (The virtual patient can simultaneously make a gesture pointing towards the lower right jaw area.)

[0070] Virtual patients can proactively supplement information about their condition or ask questions in specific situations based on the logic of the disease's development, such as, "Doctor, do I need to have a tooth extracted? I'm very afraid of tooth extractions..." This design simulates the various reactions that patients exhibit due to pain or tension in real clinical consultations, thereby enhancing learners' clinical adaptability and communication skills.

[0071] The entire process of the simulated consultation is recorded by the system, forming a complete consultation dialogue record for subsequent scoring and analysis.

[0072] S3, Medical Record Writing

[0073] After the simulated consultation, learners use the information obtained from the simulated consultation to complete the medical record writing interface within the system. The medical record writing interface is designed according to the standard format of dental clinic medical records, including sections such as chief complaint, present illness, systemic medical history, oral examination record, analysis of auxiliary examination results, preliminary diagnosis, treatment plan and treatment operation record.

[0074] Learners are required to summarize, organize, and logically reconstruct all valid information obtained during the consultation process, and write it according to medical documentation standards. The system can set a writing time limit to simulate the time pressure in actual clinical work and cultivate learners' ability to complete standardized medical record writing within a limited time.

[0075] S4, Scoring and Feedback

[0076] After the medical record is completed, the AI ​​scoring feedback module automatically scores the learner's consultation dialogue record and medical record document from multiple dimensions, and outputs a feedback report containing the scoring results and improvement suggestions in real time.

[0077] Specifically, the AI ​​scoring feedback module compares the learner's consultation records and medical records with preset scoring criteria item by item, and assigns scores and written comments for each item according to professional oral medicine standards. The scoring covers dimensions such as medical history taking, oral examination, various sub-items of medical record writing (chief complaint, present illness, systemic medical history, description of oral examination, diagnosis, treatment plan), and doctor-patient communication.

[0078] Comparison of immediate patient reception scores between two groups of resident physicians after practice

[0079]

[0080] Comparison of delayed (4 weeks later) patient admission scores between two groups of resident physicians after training

[0081]

[0082] Resident physicians used AI software to practice independently analyzing AI scoring changes during the reception of patients with different oral diseases.

[0083]

[0084]

[0085] S5. Learning Data Recording and Analysis

[0086] The learning data recording and analysis module records all process data and scoring results for this exercise, and updates the learner's personal learning progress analysis report.

[0087] In one embodiment, the system automatically generates a detailed summary report of the exercise after the course ends, including: the overall score and scores for each dimension of each exercise; a ranking of items where points are easily lost (statistics of frequently deducted points); a comparison of scores with other resident physicians of the same grade and stage; and suggestions for addressing weaknesses and targeted practice.

[0088] After multiple rounds of practice, the system integrates historical data longitudinally to create visualized learning curves, intuitively displaying the learner's progress. The system can also automatically adjust the recommended practice types and difficulty levels for the next stage based on learning progress, achieving intelligent adaptive adjustment of the learning path.

[0089] III. Application Effect Verification

[0090] To verify the practical application effect of the teaching system of this invention, a before-and-after controlled study was designed with standardized training students in general dentistry. The specific study design is as follows:

[0091] (I) Research Plan

[0092] Twenty-four resident physicians who had completed their general dental rotations were randomly divided into a control group (self-practice group, N=12) and an experimental group (AI practice group, N=12). Both groups underwent baseline theoretical examinations and self-assessments of their theoretical knowledge upon enrollment, confirming no statistically significant differences between the two groups.

[0093] Control group: The traditional self-practice method was used, and the standard case was used for free practice for 1 day.

[0094] Experimental group: Received LLM software-assisted training from the teaching system of this invention and practiced using the AI ​​virtual patient interaction module for one day.

[0095] After the practice, both groups of resident physicians underwent assessments of patient reception and medical record writing. The same group of examiners scored them using a standardized resident clinical patient reception assessment scale and recorded the time taken for each patient visit. Simultaneously, the resident physicians completed a self-assessment scale after the patient reception assessment. Four weeks later, both groups of resident physicians underwent a delayed retest to assess their ability to receive standard patients and write medical records for different diseases. Again, examiners scored and recorded the time taken. The experimental group of resident physicians also completed an additional clinical patient reception assessment feedback scale to collect feedback on their experience using the system of this invention.

[0096] (II) Experimental Data and Effect Analysis

[0097] 1. Comparison of patient assessment scores immediately after practice

[0098] After completing their training, the two groups of resident physicians underwent their first patient consultation assessment, and the scoring results are shown in Table 1.

[0099] The experimental results showed that the experimental group scored significantly better than the control group on multiple core dimensions (P<0.001):

[0100] Medical history collection dimensions: The median score in the experimental group was 12.00, and in the control group it was 9.00.

[0101] Oral examination dimensions: The mean score of the experimental group was 19.18 points, and that of the control group was 14.58 points (P=0.014).

[0102] In the doctor-patient communication dimension, the median score was 14.00 in the experimental group and 10.50 in the control group.

[0103] Total score: The median score in the experimental group was 87.00, and the median score in the control group was 64.00 (P<0.001).

[0104] Among the various sub-dimensions of medical record writing, there were significant differences in present illness history writing (experimental group 4.00 points vs. control group 2.00 points), systemic medical history writing (experimental group 2.00 points vs. control group 0.00 points), oral examination records (experimental group 6.00 points vs. control group 4.00 points), and treatment plans (experimental group 10.00 points vs. control group 6.00 points). The above data fully demonstrates that the system of this invention has a significant immediate effectiveness in improving resident physicians' medical history taking abilities, the standardization of oral examination records, treatment plan development, and doctor-patient communication skills.

[0105] 2. Comparison of scores for delayed (4-week) patient visits

[0106] To assess the learning retention effect, the two groups of resident physicians underwent a retest of their patient reception and medical record writing for different diseases 4 weeks after the initial training. The results are shown in Table 2.

[0107] Data shows that the experimental group still has a significant advantage in delayed retesting:

[0108] Medical history collection dimensions: The median score in the experimental group was 12.00, and in the control group it was 9.00 (P<0.001).

[0109] Oral examination dimensions: The median score in the experimental group was 20.00, and in the control group it was 14.50 (P=0.004).

[0110] Doctor-patient communication dimension: The median score in the experimental group was 13.00, and in the control group it was 10.50 (P=0.005).

[0111] Total score: The mean score of the experimental group was 83.10, and that of the control group was 62.92 (P<0.001).

[0112] It is noteworthy that the scores in the experimental group remained at a high level after 4 weeks, while the total score in the control group showed no significant improvement compared to the immediate assessment. This indicates that the present invention not only achieves rapid improvement in abilities in the short term, but also has a good learning retention effect, especially having a profound impact on higher-level clinical thinking abilities such as the completeness of systematic medical history collection and the rationality of treatment plan formulation.

[0113] 3. Longitudinal Change Analysis of AI Scores

[0114] To further verify the system's support effect on repeated practice, the AI ​​score changes of the resident physicians in the experimental group when using the system of this invention to practice different oral diseases (pulpitis, periodontitis, impacted teeth) for the first, second and third time were recorded. The results are shown in Table 3.

[0115] Take pulpitis as an example:

[0116] Total score for the first practice session: Median 84.00 points;

[0117] Total score for the second practice session: Median 92.00 points;

[0118] Total score of the third practice session: median 96.00 (P<0.001).

[0119] Significant improvements were observed in all core assessment dimensions: the median score for current medical history collection increased from 15.00 in the first assessment to 19.00 in the third assessment (P<0.001); the score for systemic medical history collection increased from 4.00 in the first assessment to 5.00 in the third assessment (P=0.009); and the score for oral examination increased from 20.00 in the first assessment to 23.00 in the third assessment (P=0.026).

[0120] Similar trends were observed in periodontitis cases (P=0.017) and impacted tooth cases (P=0.006), with all residents in the experimental group showing significant improvement in their total scores after three practice sessions. These results indicate that standardized interactive practice based on LLM supports a closed-loop teaching process of repeated "patient history taking—writing—scoring feedback—continuous improvement," significantly promoting the continuous enhancement of residents' comprehensive clinical skills.

[0121] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A teaching system for dental students' patient reception and medical record writing based on a large language, characterized in that, include: The case database module stores standard case data for various oral diseases. The standard case data includes at least the patient's basic information, chief complaint, present illness history, systemic medical history, and oral examination results. The AI ​​virtual patient interaction module, built on a large language model, is used to simulate a virtual standard patient based on the standard case data, generate responses that conform to the actual clinical situation in real time based on the consultation content input by the learner, and conduct simulated consultation dialogue with the learner. The AI ​​scoring feedback module, built on a large language model and preset multi-dimensional scoring standards, is used to automatically score the consultation dialogue records and subsequent medical records generated by the learner during the consultation simulation, and output the scoring results and corresponding improvement suggestions. The learning data recording and analysis module is used to automatically record the complete process data and multi-dimensional scoring results of each practice session of the learner, and analyze the data to generate a learning progress analysis report for the learner.

2. The teaching system for oral medicine students' patient reception and medical record writing based on big language as described in claim 1, characterized in that: The AI ​​virtual patient interaction module is configured to simulate standard virtual patients in the dental department of different ages, with different chief complaints, different oral diseases, and different emotional states, and dynamically adjust the response content according to the learner's consultation strategy.

3. The teaching system for oral medicine students' patient reception and medical record writing based on big language as described in claim 2, characterized in that: The AI ​​scoring feedback module is based on a multi-dimensional scoring standard that includes: medical history taking scoring dimension, oral examination scoring dimension, medical record writing scoring dimension, diagnosis and treatment plan scoring dimension, and doctor-patient communication scoring dimension.

4. The teaching system for oral medicine students' patient reception and medical record writing based on big language as described in claim 3, characterized in that: The scoring dimensions for medical record writing are further subdivided into at least three sub-dimensions: completeness of chief complaint, logicality of present illness history, completeness of systemic medical history, standardization of oral examination description, accuracy of diagnosis, and rationality of treatment plan.

5. The teaching system for oral medicine students' patient reception and medical record writing based on big language as described in claim 4, characterized in that: The AI ​​scoring feedback module is also configured to provide correct examples and improvement suggestions for each deducted point while generating the scoring results.

6. The teaching system for oral medicine students' patient reception and medical record writing based on big language as described in claim 5, characterized in that: The learning data recording and analysis module is also configured to automatically identify the learner's weak knowledge areas and ability improvement trends through longitudinal comparative analysis of the practice data, and generate a visual learning curve report.

7. A teaching method for dental students' patient reception and medical record writing, applied to the teaching system for dental students' patient reception and medical record writing based on big language as described in any one of claims 1 to 6, characterized in that, Includes the following steps: S1. Case push: Select or push standard cases of the target oral disease from the case database to the learner; S2, AI-assisted patient reception simulation: The learner, based on the pushed standard case information, conducts real-time intelligent dialogue with a virtual standard patient through the AI ​​virtual patient interaction module to carry out simulated practice of consultation and medical history collection; S3. Medical Record Writing: After the simulated consultation, the learner completes the medical record writing within the system based on the information obtained from the simulated consultation. S4. Scoring and Feedback: The AI ​​scoring and feedback module automatically scores the learner's consultation dialogue records and medical records from multiple dimensions, and outputs a feedback report containing the scoring results and improvement suggestions in real time. S5. Learning Data Recording and Analysis: The learning data recording and analysis module records all process data and scoring results of this exercise, and updates the learner's personal learning progress analysis report.

8. The teaching method of the teaching system for oral medicine students' patient reception and medical record writing based on big language as described in claim 7, characterized in that: In step S2, the AI ​​virtual patient interaction module generates a response by intelligently analyzing and matching answers to the questions input by the learner, based on the preset medical condition information in the standard case data and combined with the oral medicine professional knowledge base and clinical diagnosis and treatment logic.

9. The teaching method of the teaching system for oral medicine students' patient reception and medical record writing based on big language as described in claim 8, characterized in that: In step S4, the AI ​​scoring feedback module compares the learner's consultation dialogue records and medical records with the preset scoring points one by one, and gives scores and text comments for each item according to the professional standards of oral medicine.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the teaching method according to any one of claims 7 to 9.

Images