Method and device for improving doctor inquiry ability based on virtual patient

Abstract

The invention relates to the technical field of large models, and particularly provides a method and device for improving doctor inquiry ability based on a virtual patient, and the method comprises the following steps: S1, constructing the virtual patient; s2, performing inquiry process simulation; and S3, evaluating the inquiry structure. Compared with the prior art, an efficient and convenient inquiry training and examination platform can be provided for doctors, and the inquiry ability of the doctors can be improved.

Description

Technical Field

[0001] The present invention relates to the field of large model technology, and specifically provides a method and device for improving doctors' diagnosis capabilities based on virtual patients. Background Art

[0002] In medical education and physician training, the ability to conduct a clinical interview is a crucial skill for doctors. Traditional clinical interview training often relies on simulations with real patients or standardized patients. However, these methods have numerous limitations, including limited access to real patients, high costs for standardized patient training, and difficulty covering all disease types and scenarios. Furthermore, objectively and accurately assessing a doctor's clinical interview skills has always been a challenge in physician examinations.

[0003] In recent years, with the rapid development of artificial intelligence (AI), large language models (LLMs) have made significant progress in natural language processing, demonstrating powerful language generation and comprehension capabilities. LLMs, when used to construct virtual patients, offer a method for improving physicians' consultation capabilities and enhancing the efficiency of physician consultation testing. Summary of the Invention

[0004] The present invention aims to address the deficiencies of the above-mentioned prior art and provides a highly practical method for improving doctors' diagnosis capabilities based on virtual patients.

[0005] A further technical task of the present invention is to provide a device that is rationally designed, safe and applicable and can enhance the doctor's diagnosis ability based on virtual patients.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] A method for improving a doctor's consultation ability based on a virtual patient comprises the following steps:

[0008] S1, virtual patient construction;

[0009] S2, simulation of the consultation process;

[0010] S3. Evaluation of the interview structure.

[0011] Furthermore, in step S1, it includes:

[0012] S1-1, basic model selection and training;

[0013] S1-2, virtual patient design;

[0014] S1-3. Diagnosis and treatment process design.

[0015] Furthermore, in step S1-1, it includes:

[0016] S1-1.1. Select an open-source large language model as the base model and fine-tune it to incorporate medical expertise and language expression capabilities.

[0017] S1-1.2. Preprocess the collected medical record data and input the processed medical record data into the large language model for training, so that the model can generate corresponding answers based on different medical questions;

[0018] S1-1.3. During the training process, supervised learning methods are used to train the model through labeled medical questions and answers to optimize the model parameters.

[0019] Furthermore, in step S1-2, it includes:

[0020] S1-2.1. Design the virtual identity and background information of the virtual patient;

[0021] S1-2.2. Combine the virtual patient's medical records with the large language model, so that the virtual patient can respond to the doctor's questions based on their own medical records and the model's generation capabilities;

[0022] In step S1-3, it includes:

[0023] S1-3.1. Identify the patient's role in the consultation process;

[0024] S1-3.2. Clarify the logical relationship between the key links in the patient consultation process.

[0025] Furthermore, in step S2, it includes:

[0026] S2-1. After logging into the system, the doctor selects a virtual patient and generates the initial state and answers of the virtual patient based on the identity and background information of the virtual patient;

[0027] S2-2, based on the virtual patient selected by the doctor, loading the identity and background information of the virtual patient, and generating the initial state and answer of the virtual patient;

[0028] S2-3. The doctor starts the consultation by greeting the virtual patient and asking about their purpose. The virtual patient responds according to the preset answer pattern. The doctor asks further questions based on the virtual patient's answers. The virtual patient responds accordingly based on the doctor's questions, combined with their own medical records and the generation capabilities of the large language model.

[0029] S2-4. During the consultation process, the doctor's questions and the virtual patient's answers are recorded in real time to form a complete consultation dialogue record.

[0030] Furthermore, in step S3, it includes:

[0031] S3-1. Use natural language processing technology to summarize the consultation process and obtain the patient information collected during this process;

[0032] S3-2. Compare the virtual patient information provided by the system with the results of the current consultation and evaluate the doctor's consultation ability from multiple dimensions of the consultation;

[0033] S3-3. Feedback the scoring results to the doctor and provide a detailed evaluation report, which points out the doctor's strengths and weaknesses during the consultation process.

[0034] Furthermore, in step S3-2, it includes:

[0035] S3-2.1, completeness of medical consultation: score based on whether the doctor inquired about the chief complaint, current medical history, past medical history, family history, and marital and reproductive history, with a maximum score of 8 points for each item and a total score of 40 points;

[0036] S3-2.2, Questioning Accuracy, is scored based on whether the questions asked by the doctor are accurate and targeted, and whether the answers of the virtual patient are consistent with the preset case information, with a maximum score of 40 points;

[0037] S3-2.3. Score the doctor's logical and organized interview process, with a maximum score of 10 points;

[0038] S3-2.4: Score the doctor's communication style based on whether he or she is appropriate and whether he or she can guide the patient to answer questions, with a maximum score of 10 points;

[0039] S3-2.5. Score the doctor's consultation process according to the scoring mechanism to generate a score for this consultation.

[0040] A device for improving a doctor's consultation ability based on a virtual patient, characterized by comprising: at least one memory and at least one processor;

[0041] The at least one memory is configured to store a machine-readable program;

[0042] The at least one processor is configured to call the machine-readable program to execute a method for improving a doctor's diagnosis capability based on a virtual patient.

[0043] Compared with the prior art, the method and device of the present invention for improving doctors' consultation ability based on virtual patients have the following outstanding beneficial effects:

[0044] (1) The present invention constructs a virtual patient system based on large model technology to simulate real consultation scenarios, providing doctors with an efficient and convenient consultation training and examination platform, which helps to improve doctors' consultation capabilities.

[0045] (2) The virtual patient system can generate accurate and reasonable answers based on the doctor's questions and cover all key links in the consultation process, thereby improving the authenticity and comprehensiveness of the consultation training.

[0046] (3) The scientific and objective scoring mechanism designed can accurately evaluate the doctor's consultation process, providing an efficient and objective consultation ability assessment tool for the physician examination, which helps to improve the quality and fairness of the physician examination.

[0047] (4) The virtual patient system of the present invention is scalable and flexible, can be customized and optimized according to different medical fields and examination requirements, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] 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.

[0049] Attachment Figure 1 It is a flowchart of a method for improving doctors' consultation ability based on virtual patients;

[0050] Attachment Figure 2 It is a flowchart of a virtual patient construction method for improving a doctor's consultation ability based on virtual patients;

[0051] Attachment Figure 3 It is a flowchart of simulating the consultation process in a method for improving the doctor's consultation ability based on virtual patients;

[0052] Attachment Figure 4 The present invention is a flowchart of the evaluation of consultation results in a method for improving the doctor's consultation ability based on virtual patients. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to better understand the solutions of the present invention, the present invention will be further described in detail below in conjunction with specific embodiments. Obviously, the embodiments described are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0054] A best embodiment is given below:

[0055] like Figure 1As shown, a method for improving doctors' consultation ability based on virtual patients in this embodiment has the following steps:

[0056] S1, virtual patient construction;

[0057] like Figure 2 As shown, including:

[0058] S1-1, basic model selection and training;

[0059] include:

[0060] S1-1.1. Select an open-source large language model (such as GPT-3 or BERT) as the base model and fine-tune it to incorporate medical expertise and language expression capabilities. During fine-tuning, the model is trained using a large amount of medical literature and medical records to enable it to understand and generate medical-related language content.

[0061] S1-1.2. Preprocess the collected medical record data, including data cleaning and annotation, to ensure data quality and usability. The processed medical record data is then fed into a large language model for training, enabling the model to generate appropriate responses based on different medical questions.

[0062] S1-1.3. During the training process, supervised learning methods are used to train the model through labeled medical questions and answers, optimize the model parameters, and improve the generation quality and accuracy of the model.

[0063] S1-2, virtual patient design;

[0064] include:

[0065] S1-2.1. Design the virtual patient's identity and background information, including age, gender, occupation, personality traits, etc. For example, create a 35-year-old male patient who is a programmer and has an introverted personality, etc., to make the virtual patient more realistic and personalized.

[0066] S1-2.2. Combine the medical records of virtual patients with the large language model so that the virtual patients can give corresponding answers based on the doctor's questions, combined with their own medical records and the generation capabilities of the model.

[0067] S1-3, diagnosis and treatment process design;

[0068] include:

[0069] S1-3.1. Identify the steps involved in the patient's medical consultation, such as chief complaint, current medical history, past medical history, family history, reproductive history, physical examination, and auxiliary examinations.

[0070] S1-3.2. Clarify the logical relationships between the key steps in the patient consultation process. For example, build a complete consultation process framework from the chief complaint to the present medical history, then to the past medical history and family history.

[0071] S2, simulation of the consultation process;

[0072] Used to simulate the consultation dialogue process between doctors and virtual patients, including the interaction between doctors and virtual patients, real-time recognition and recording of consultation dialogues, etc.

[0073] like Figure 3 Shown, including:

[0074] S2-1. After the doctor logs into the system, he selects a virtual patient. The system generates the initial state and answers of the virtual patient based on the identity and background information of the virtual patient.

[0075] S2-2. The system loads the identity and background information of the virtual patient selected by the doctor and generates the initial state and answer of the virtual patient.

[0076] S2-3. The doctor starts the consultation dialogue by greeting the virtual patient and asking about his purpose of visit.

[0077] The virtual patient responds according to a pre-set answer pattern. Based on the virtual patient's answers, the doctor further inquires about the patient's chief complaint, current medical history, past medical history, marital and reproductive history, family history, physical examination results, and auxiliary examinations. The virtual patient responds accordingly based on the doctor's questions, combining their own medical records and the generation capabilities of the large language model.

[0078] S2-4. During the consultation process, the system records the doctor's questions and the virtual patient's answers in real time, forming a complete consultation dialogue record.

[0079] S3, Questioning Structure Assessment;

[0080] After the consultation is completed, the system summarizes and analyzes the doctor's consultation process based on the virtual patient's answers and the pre-stored medical record information of the virtual patient.

[0081] like Figure 4 Shown, including:

[0082] S3-1. The system uses natural language processing technology to summarize the consultation process and obtain the patient information collected during this process, including: chief complaint, current medical history, past medical history, family history, and marital and reproductive history.

[0083] S3-2. Compare the virtual patient information provided by the system with the results of the current consultation, and evaluate the doctor's consultation ability from multiple dimensions such as the completeness, accuracy, logic, and communication skills of the consultation.

[0084] For example:

[0085] (1) Completeness of medical consultation: The score is calculated based on whether the doctor has asked about key information such as the chief complaint, current medical history, past medical history, family history, and marital and reproductive history. The maximum score for each item is 8 points, and the total score is 40 points.

[0086] (2) Interview accuracy: The accuracy of the questions asked by the doctor is accurate and targeted, and the answers of the virtual patient are consistent with the preset case information. The maximum score is 40 points.

[0087] (3) Logic of medical consultation: The doctor’s medical consultation process is scored based on whether it is logical and organized, with a maximum score of 10 points.

[0088] (4) Communication skills: The doctor is scored based on whether his communication style is appropriate and whether he can guide the patient to answer questions, with a maximum score of 10 points.

[0089] The system scores the doctor's consultation process according to the scoring mechanism to generate a score for this consultation. For example, the doctor's consultation score is: 40 points for consultation completeness, 40 points for consultation accuracy, 10 points for consultation logic, and 10 points for communication skills, for a total score of 100 points.

[0090] S3-3. The system provides the doctor with feedback on the scoring results and a detailed evaluation report. The evaluation report identifies the doctor's strengths and weaknesses during the consultation process. For example, "The doctor was able to inquire about the patient's chief complaint and current medical history relatively thoroughly during the consultation, but omitted information on family history and marital and reproductive history; the consultation accuracy was high, but some questions were not specific enough; the consultation logic was good, but communication skills need further improvement." The doctor can use the evaluation report to understand their own consultation level and conduct targeted training and improvement to address any deficiencies.

[0091] Based on the above method, a device for improving a doctor's consultation ability based on a virtual patient in this embodiment is characterized by comprising: at least one memory and at least one processor;

[0092] The at least one memory is configured to store a machine-readable program;

[0093] The at least one processor is configured to call the machine-readable program to execute a method for improving a doctor's diagnosis capability based on a virtual patient.

[0094] The above-mentioned specific implementation methods are only specific cases of the present invention. The patent protection scope of the present invention includes but is not limited to the above-mentioned specific implementation methods. Any technical solutions that conform to the above-mentioned specific implementation methods of the present invention and any appropriate changes or substitutions made thereto by ordinary technicians in the relevant technical field shall fall within the patent protection scope of the present invention.

[0095] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for improving doctors' consultation ability based on virtual patients, characterized in that: The steps are as follows: S1, virtual patient construction; S2, simulation of the consultation process; S3. Evaluation of the interview structure.

2. The method for improving doctors' consultation ability based on virtual patients according to claim 1, characterized in that: In step S1, it includes: S1-1, basic model selection and training; S1-2, virtual patient design; S1-3. Diagnosis and treatment process design.

3. The method for improving doctors' consultation ability based on virtual patients according to claim 2, characterized in that: In step S1-1, it includes: S1-1.

1. Select an open-source large language model as the base model and fine-tune it to incorporate medical expertise and language expression capabilities. S1-1.

2. Preprocess the collected medical record data and input the processed medical record data into the large language model for training, so that the model can generate corresponding answers based on different medical questions; S1-1.

3. During the training process, supervised learning methods are used to train the model through labeled medical questions and answers to optimize the model parameters.

4. The method for improving doctors' consultation ability based on virtual patients according to claim 3, characterized in that: In step S1-2, it includes: S1-2.

1. Design the virtual identity and background information of the virtual patient; S1-2.

2. Combine the virtual patient's medical records with the large language model, so that the virtual patient can respond to the doctor's questions based on their own medical records and the model's generation capabilities; In step S1-3, it includes: S1-3.

1. Identify the patient's role in the consultation process; S1-3.

2. Clarify the logical relationship between the key links in the patient consultation process.

5. The method for improving doctors' consultation ability based on virtual patients according to claim 4, characterized in that: In step S2, it includes: S2-1. After logging into the system, the doctor selects a virtual patient and generates the initial state and answers of the virtual patient based on the identity and background information of the virtual patient; S2-2, based on the virtual patient selected by the doctor, loading the identity and background information of the virtual patient, and generating the initial state and answer of the virtual patient; S2-3. The doctor starts the consultation by greeting the virtual patient and asking about their purpose. The virtual patient responds according to the preset answer pattern. The doctor asks further questions based on the virtual patient's answers. The virtual patient responds accordingly based on the doctor's questions, combined with their own medical records and the generation capabilities of the large language model. S2-4. During the consultation process, the doctor's questions and the virtual patient's answers are recorded in real time to form a complete consultation dialogue record.

6. The method for improving doctors' consultation ability based on virtual patients according to claim 5, characterized in that: In step S3, it includes: S3-1. Use natural language processing technology to summarize the consultation process and obtain the patient information collected during this process; S3-2. Compare the virtual patient information provided by the system with the results of the current consultation and evaluate the doctor's consultation ability from multiple dimensions of the consultation; S3-3. Feedback the scoring results to the doctor and provide a detailed evaluation report, which points out the doctor's strengths and weaknesses during the consultation process.

7. The method for improving doctors' consultation ability based on virtual patients according to claim 6, characterized in that: In step S3-2, it includes: S3-2.1, completeness of medical consultation: score based on whether the doctor inquired about the chief complaint, current medical history, past medical history, family history, and marital and reproductive history, with a maximum score of 8 points for each item and a total score of 40 points; S3-2.2, Questioning Accuracy, is scored based on whether the questions asked by the doctor are accurate and targeted, and whether the answers of the virtual patient are consistent with the preset case information, with a maximum score of 40 points; S3-2.

3. Score the doctor's logical and organized interview process, with a maximum score of 10 points; S3-2.4: Score the doctor's communication style based on whether he or she is appropriate and whether he or she can guide the patient to answer questions, with a maximum score of 10 points; S3-2.

5. Score the doctor's consultation process according to the scoring mechanism to generate a score for this consultation.

8. A device for improving doctors' consultation ability based on virtual patients, characterized by: include: at least one memory and at least one processor; The at least one memory is configured to store a machine-readable program; The at least one processor is configured to call the machine-readable program to execute the method according to any one of claims 1 to 7.

Images