Intelligent diagnosis method, device, storage medium and equipment for patient medical history

Abstract

The present application relates to the field of smart medical technology and discloses a method, apparatus, storage medium, and computer device for intelligent medical consultation based on a patient's medical history. The method comprises: in response to a request for an intelligent medical consultation service, outputting a first round of doctor conversation information with an intelligent doctor, obtaining a first round of user conversation information for the first round of doctor conversation information, and identifying medical history keywords in the first round of user conversation information; based on the identified medical history keywords, obtaining multiple medical inquiry points corresponding to the medical history keywords from a preset structured knowledge base as medical inquiry key points information, wherein the preset structured knowledge base contains multiple medical inquiry key points corresponding to multiple preset medical history keywords; generating doctor conversation information based on the first round of user conversation information and the medical inquiry key points information and outputting the information until the conversation between the target user and the intelligent doctor covers all the medical inquiry key points information. This improves the efficiency and accuracy of medical consultation.

Description

Technical Field

[0001] The present application relates to the field of smart medical technology, and in particular to an intelligent medical consultation method, device, storage medium and computer equipment for a patient's medical history. Background Art

[0002] In modern healthcare, accurate patient medical history information is crucial for doctors to make correct diagnoses. Traditionally, the consultation process relies on face-to-face interaction between doctors and patients, where doctors gather patient history through questioning and listening. However, this approach presents several challenges.

[0003] First, face-to-face consultations are limited by time and space. Doctors need to obtain as much information as possible from patients within a limited timeframe, which can lead to the omission of key information. Second, different doctors may question patients based on their own experience and habits, resulting in differences in the content of the medical history, which can hinder subsequent diagnosis and treatment.

[0004] To address these issues, computer-assisted consultation systems have emerged in recent years. These systems typically use pre-set consultation templates to guide doctors through the consultation process, thereby improving the standardization of consultations. However, these systems still have some shortcomings. For example, they often only conduct consultations according to fixed templates and cannot be flexibly adjusted to the patient's specific situation. Summary of the Invention

[0005] In view of this, the embodiments of the present application provide an intelligent medical consultation method, apparatus, storage medium and computer equipment for a patient's medical history, which improves the accuracy and efficiency of the medical consultation.

[0006] According to one aspect of the present application, a method for intelligent diagnosis of a patient's medical history is provided, the method comprising:

[0007] In response to a request for the intelligent consultation service, outputting first-round doctor conversation information of the intelligent doctor, obtaining first-round user conversation information for the first-round doctor conversation information, and performing medical history keyword recognition on the first-round user conversation information;

[0008] Based on the identified medical history keywords, a plurality of medical inquiry key points corresponding to the medical history keywords are obtained from a preset structured knowledge base as medical inquiry key point information, wherein the preset structured knowledge base contains a plurality of medical inquiry key points corresponding to each of the plurality of preset medical history keywords;

[0009] Based on the first round of user dialogue information and the medical inquiry key points information, doctor dialogue information is generated and outputted until the dialogue between the target user and the intelligent doctor covers all medical inquiry key points information.

[0010] Optionally, based on the identified medical history keywords, obtaining a plurality of medical inquiry key points corresponding to the medical history keywords in a preset structured knowledge base as medical inquiry key point information includes:

[0011] Obtain the consultation category of the request initiation instruction;

[0012] A plurality of medical consultation points corresponding to the disease keywords under the consultation category are obtained from the preset structured knowledge base as the medical consultation point information.

[0013] Optionally, generating and outputting doctor dialogue information based on the first round of user dialogue information and the consultation key points information includes:

[0014] The first round of doctor conversation information and the first round of user conversation information are used as current round conversation information. The current round conversation information is subjected to a key point understanding model based on the key point information to determine whether the current round conversation information satisfies a key point information obtaining condition for any key point in the key point information, wherein the key point information obtaining condition includes that the key point has been asked and an answer that complies with a rule has been obtained.

[0015] If the key point information obtaining condition of any medical consultation point is met, the satisfied medical consultation point will be removed from the medical consultation point information, and if the medical consultation point information is not cleared, the next round of doctor dialogue information will continue to be generated; otherwise, the next round of doctor dialogue information will stop being generated;

[0016] If the key information acquisition conditions of any of the medical consultation points are not met, the next round of doctor dialogue information is directly generated.

[0017] Optionally, before understanding the key points of the current round of dialogue information based on the key points of the medical inquiry using the large model for understanding the key points of the medical inquiry, the method further includes:

[0018] Obtain a sample of medical consultation dialogue information pre-labeled with a medical consultation key point label, wherein the medical consultation dialogue information sample includes at least one round of medical consultation dialogue, each round of medical consultation dialogue is labeled with a medical consultation key point label, and the medical consultation key point label includes an empty label and multiple medical consultation key point identification labels, wherein the empty label indicates that the corresponding round of medical consultation dialogue does not contain medical consultation key points or does not meet the key point information acquisition conditions, and the medical consultation key point identification label indicates the identification of the medical consultation key points contained in the corresponding round of medical consultation dialogue;

[0019] The initial large model is fine-tuned and trained based on the medical consultation dialogue information sample to obtain the large model for understanding the key points of the medical consultation.

[0020] Optionally, after removing the satisfied medical inquiry points from the medical inquiry points information, the method further includes:

[0021] The removed key points are recorded as key points asked;

[0022] Generate the next round of doctor dialogue information, including:

[0023] Generate multiple candidate doctor dialogue messages for the current round of user dialogue information based on the current round of dialogue information and historical dialogue context through the consultation dialogue big model;

[0024] Identify the key points of the candidate doctor conversation information, delete the candidate doctor conversation information containing the key points of the question, select one of the remaining candidate doctor conversation information as the next round of doctor conversation information for output, and use the next round of doctor conversation information and the user conversation information corresponding to the next round of doctor conversation information as the new current round conversation information, and return to the step of understanding the key points of the current round conversation information based on the key points of the question.

[0025] Optionally, the large model of the medical consultation dialogue generates, based on the current round of dialogue information and the historical dialogue context, multiple candidate doctor dialogue information of the intelligent doctor for the current round of user dialogue information, including:

[0026] Determine whether the current conversation round has reached the preset conversation round;

[0027] If the current conversation round does not reach the preset conversation round, the medical consultation conversation model is used to generate multiple candidate doctor conversation information for the current round of user conversation information based on the current round of conversation information and historical conversation context;

[0028] If the current conversation round has reached the preset conversation round, a conversation generation prompt word is constructed based on the medical inquiry key point information, the current round conversation information and the historical conversation context, and the conversation generation prompt word is input into the medical inquiry conversation model, so that the medical inquiry conversation model generates the next round of doctor conversation information containing any medical inquiry key point in the medical inquiry key point information based on the current round conversation information and the historical conversation context.

[0029] Optionally, selecting one piece of the remaining candidate doctor dialogue information as the next round of doctor dialogue information for output includes:

[0030] If any of the remaining candidate doctor dialogue information contains any of the medical inquiry points in the medical inquiry key points information, then one of the candidate doctor dialogue information containing any of the medical inquiry key points is selected as the next round of doctor dialogue information for output;

[0031] If the remaining candidate doctor dialogue information does not contain any of the medical inquiry points in the medical inquiry key points information, then one piece of the remaining candidate doctor dialogue information is selected as the next round of doctor dialogue information for output.

[0032] According to another aspect of the present application, there is provided an intelligent medical history inquiry device for a patient, the device comprising:

[0033] a keyword recognition module configured to, in response to a request for the intelligent consultation service, output a first round of doctor conversation information of the intelligent doctor, obtain a first round of user conversation information for the first round of doctor conversation information, and perform medical history keyword recognition on the first round of user conversation information;

[0034] A medical question key point acquisition module is used to obtain, based on the identified medical history keywords, a plurality of medical question key points corresponding to the medical history keywords from a preset structured knowledge base as medical question key point information, wherein the preset structured knowledge base contains a plurality of medical question key points corresponding to each of the plurality of preset medical history keywords;

[0035] The intelligent consultation module is used to generate doctor dialogue information based on the first round of user dialogue information and the consultation key points information for output until the dialogue between the target user and the intelligent doctor covers all the consultation key points information.

[0036] Optionally, the medical consultation key points acquisition module is further used to:

[0037] Obtain the consultation category of the request initiation instruction;

[0038] A plurality of medical consultation points corresponding to the disease keywords under the consultation category are obtained from the preset structured knowledge base as the medical consultation point information.

[0039] Optionally, the intelligent medical consultation module is further used to:

[0040] The first round of doctor conversation information and the first round of user conversation information are used as current round conversation information. The current round conversation information is subjected to a key point understanding model based on the key point information to determine whether the current round conversation information satisfies a key point information obtaining condition for any key point in the key point information, wherein the key point information obtaining condition includes that the key point has been asked and an answer that complies with a rule has been obtained.

[0041] If the key point information obtaining condition of any medical consultation point is met, the satisfied medical consultation point will be removed from the medical consultation point information, and if the medical consultation point information is not cleared, the next round of doctor dialogue information will continue to be generated; otherwise, the next round of doctor dialogue information will stop being generated;

[0042] If the key information acquisition conditions of any of the medical consultation points are not met, the next round of doctor dialogue information is directly generated.

[0043] Optionally, the device further includes a model training module, configured to:

[0044] Obtain a sample of medical consultation dialogue information pre-labeled with a medical consultation key point label, wherein the medical consultation dialogue information sample includes at least one round of medical consultation dialogue, each round of medical consultation dialogue is labeled with a medical consultation key point label, and the medical consultation key point label includes an empty label and multiple medical consultation key point identification labels, wherein the empty label indicates that the corresponding round of medical consultation dialogue does not contain medical consultation key points or does not meet the key point information acquisition conditions, and the medical consultation key point identification label indicates the identification of the medical consultation key points contained in the corresponding round of medical consultation dialogue;

[0045] The initial large model is fine-tuned and trained based on the medical consultation dialogue information sample to obtain the large model for understanding the key points of the medical consultation.

[0046] Optionally, the intelligent medical consultation module is further used to:

[0047] The removed key points are recorded as key points asked;

[0048] Through the large model of medical consultation dialogue, based on the current round of dialogue information and historical dialogue context, the intelligent doctor generates multiple candidate doctor dialogue information for the current round of user dialogue information; the medical consultation key points of the candidate doctor dialogue information are identified, the candidate doctor dialogue information containing the medical consultation key points are deleted, and one of the remaining candidate doctor dialogue information is selected as the next round of doctor dialogue information for output, and the next round of doctor dialogue information and the user dialogue information corresponding to the next round of doctor dialogue information are used as the new current round of dialogue information, and the process returns to the step of understanding the medical consultation key points of the current round of dialogue information based on the medical consultation key points information.

[0049] Optionally, the intelligent medical consultation module is further used to:

[0050] Determine whether the current conversation round has reached the preset conversation round;

[0051] If the current conversation round does not reach the preset conversation round, the medical consultation conversation model is used to generate multiple candidate doctor conversation information for the current round of user conversation information based on the current round of conversation information and historical conversation context;

[0052] If the current conversation round has reached the preset conversation round, a conversation generation prompt word is constructed based on the medical inquiry key point information, the current round conversation information and the historical conversation context, and the conversation generation prompt word is input into the medical inquiry conversation model, so that the medical inquiry conversation model generates the next round of doctor conversation information containing any medical inquiry key point in the medical inquiry key point information based on the current round conversation information and the historical conversation context.

[0053] Optionally, the intelligent medical consultation module is further used to:

[0054] If any of the remaining candidate doctor dialogue information contains any of the medical inquiry points in the medical inquiry key points information, then one of the candidate doctor dialogue information containing any of the medical inquiry key points is selected as the next round of doctor dialogue information for output;

[0055] If the remaining candidate doctor dialogue information does not contain any of the medical inquiry points in the medical inquiry key points information, then one piece of the remaining candidate doctor dialogue information is selected as the next round of doctor dialogue information for output.

[0056] According to another aspect of the present application, a storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the above-mentioned intelligent diagnosis method for the patient's medical history is implemented.

[0057] According to another aspect of the present application, a computer device is provided, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the above-mentioned intelligent diagnosis method for a patient's medical history when executing the program.

[0058] By means of the above technical solution, the embodiments of the present application provide an intelligent medical consultation method, device, storage medium, and computer equipment for a patient's medical history. Through the intelligent medical consultation service, patients can initiate medical consultation requests at any time and any place without having to wait for a face-to-face consultation with a doctor, saving patients' time and energy while also improving the efficiency of medical resource utilization. By identifying medical history keywords in user conversations and searching for relevant medical consultation key points in the knowledge base based on these keywords to guide the medical consultation process, the standardization of medical consultations is ensured, enabling the intelligent medical consultation process to more deeply explore and analyze the user's medical history information, thereby providing more accurate and comprehensive diagnostic recommendations.

[0059] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0061] Figure 1 A schematic diagram of a process for an intelligent diagnosis method for a patient's medical history provided in an embodiment of the present application is shown;

[0062] Figure 2 A flow chart of another intelligent diagnosis method for a patient's medical history provided in an embodiment of the present application is shown;

[0063] Figure 3 A schematic structural diagram of an intelligent medical history inquiry device for a patient provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0064] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of the present application can be combined with each other.

[0065] In this embodiment, an intelligent diagnosis method for a patient's medical history is provided. Figure 1 As shown, the method includes:

[0066] Step 101: In response to a request for an intelligent medical consultation service, output the first round of doctor dialogue information of the intelligent doctor, obtain the first round of user dialogue information for the first round of doctor dialogue information, and perform medical history keyword recognition on the first round of user dialogue information.

[0067] Step 102: Based on the identified medical history keywords, obtain multiple medical inquiry points corresponding to the medical history keywords in a preset structured knowledge base as medical inquiry key points information, wherein the preset structured knowledge base contains multiple medical inquiry key points corresponding to multiple preset medical history keywords.

[0068] Step 103: Generate doctor dialogue information based on the first round of user dialogue information and the consultation key points information for output until the dialogue between the target user and the intelligent doctor covers all the consultation key points information.

[0069] In an embodiment of the present application, when a user (a patient or their family member, etc.) initiates a request for the intelligent medical consultation service through some means (such as clicking a webpage button or using a mobile application), the intelligent doctor first responds to this request and outputs the first round of conversation information with the doctor. This is typically a guiding speech that guides the user to begin describing their medical history. Next, the user's first round of conversation information in response to this guiding speech is obtained and medical history keywords are identified in this information. Medical history keywords are key words that reflect the user's medical condition, symptoms, and medical history. Based on these medical history keywords, the disease name can be determined. After identifying the medical history keywords, the medical history keywords are used to query a preset structured knowledge base. This knowledge base is pre-constructed and contains multiple medical consultation points corresponding to each of the preset medical history keywords. Specifically, the preset structured knowledge base may contain multiple preset medical history keywords corresponding to each of the disease names, as well as multiple medical consultation points corresponding to each of the multiple disease names. The medical history keywords are matched with the preset medical history keywords, and the multiple medical consultation points corresponding to the disease names that match the preset medical history keywords are used as the medical consultation points information. Key points of the consultation are questions or guiding statements that target a specific medical history keyword to further provoke the user into obtaining more detailed medical history information. After obtaining these key points, the doctor generates the next round of dialogue information based on this information and the user's first round of conversation. This information is then output to the user and the doctor awaits the user's next round of conversation. This process is repeated until all key points of the consultation related to the current medical history keyword have been covered through the conversation.

[0070] In addition, it should be noted that if the medical history keywords are not identified from the first round of user dialogue information, the guiding discourse can be regenerated and output based on the first round of user dialogue information until the disease keywords can be obtained from the user dialogue information, and then the subsequent steps are executed.

[0071] By applying the technical solution of this embodiment, the intelligent consultation service allows patients to initiate consultation requests at any time and anywhere, without having to wait for a face-to-face consultation with a doctor. This saves patients time and energy, while also improving the efficiency of medical resource utilization. By identifying medical history keywords in user conversations and searching the knowledge base for relevant consultation points based on these keywords to guide the consultation process, the standardization of consultations is ensured, enabling the intelligent consultation process to more deeply explore and analyze the user's medical history information, thereby providing more accurate and comprehensive diagnostic recommendations.

[0072] Furthermore, as a refinement and expansion of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, another intelligent diagnosis method for the patient's medical history is provided, such as Figure 2 As shown, the method includes:

[0073] Step 201: In response to a request for initiating an intelligent medical consultation service, output the first round of doctor dialogue information of the intelligent doctor, obtain the first round of user dialogue information for the first round of doctor dialogue information, and perform medical history keyword recognition on the first round of user dialogue information.

[0074] In an embodiment of the present application, when a user initiates a request for the intelligent medical consultation service, the first round of dialogue information of the intelligent doctor is output. This is usually an opening statement or guiding question, intended to start the conversation. Subsequently, the user's response, that is, the first round of user dialogue information, is obtained and medical history keywords are identified. Medical history keywords are important disease or symptom information that may be involved in the user's answer. By quickly responding and guiding the user to provide medical history information, the foundation is laid for the subsequent medical consultation process and the efficiency of the consultation is improved.

[0075] Step 202: Obtain the consultation category indicated by the initiation request; obtain multiple consultation points corresponding to the disease keywords under the consultation category in the preset structured knowledge base as the consultation point information.

[0076] In an embodiment of the present application, based on the consultation category (such as depression) specified by the user when initiating the request, the key points of the consultation corresponding to the consultation category and the identified disease keywords (such as epilepsy) are searched in the preset structured knowledge base. These key points of the consultation are the key information points that the doctor needs to obtain during the consultation process. By accurately matching the consultation category and the disease keywords, the system can customize the extraction of the key points of the consultation for the user's specific situation, ensuring the pertinence and comprehensiveness of the consultation.

[0077] Step 203: Use the first round of doctor dialogue information and the first round of user dialogue information as the current round of dialogue information, and use the large model for understanding the key points of the medical consultation to understand the key points of the current round of dialogue information based on the key points of the medical consultation information to determine whether the current round of dialogue information meets the key point information acquisition conditions of any medical consultation key point in the medical consultation key point information, wherein the key point information acquisition conditions include that the medical consultation key point has been asked and an answer that complies with the rules has been obtained; if the key point information acquisition conditions of any medical consultation key point are not met, execute step 205.

[0078] In this embodiment of the present application, the first round of doctor conversation information and the first round of user conversation information are used as the current round of conversation information. This information is processed by the large model for understanding the key points of the medical consultation to determine whether the key point information acquisition conditions for any medical consultation point are met, including whether the medical consultation point has been asked and the user has given a rule-compliant answer. A rule-compliant answer usually means that the user has given a clear answer to the medical consultation point. Through the intelligent understanding of the large model, the progress of the medical consultation can be evaluated in real time to ensure that each medical consultation point receives appropriate attention and processing, avoiding omissions.

[0079] Step 204: If the conditions for obtaining the key information of any medical consultation point are met, the satisfied medical consultation point will be removed from the medical consultation point information, and the removed medical consultation point will be recorded as the medical consultation point that has been asked. If the medical consultation point information has not been cleared, step 205 will be executed, otherwise the generation of the next round of doctor dialogue information will be stopped.

[0080] In an embodiment of the present application, if it is found that a certain inquiry point has met the conditions for obtaining the key point information, it will be removed from the inquiry point information and marked as the inquiry point that has been asked. If all the inquiry points have been met, the generation of the next round of doctor dialogue information will be stopped, because the necessary inquiry process has been completed at this time, and the subsequent process can be continued, such as sending the inquiry record to the doctor terminal for evaluation by the doctor terminal. By dynamically updating the inquiry point list, the system can efficiently manage the inquiry process, ensure the completeness and accuracy of the inquiry, and avoid unnecessary repeated inquiries. In addition, when it is determined to stop generating the next round of doctor dialogue information, a dialogue record can also be generated based on the dialogue information of each round. The dialogue record marks the inquiry points contained in each round of dialogue and the answers given by the user to the inquiry points that comply with the rules, so that subsequent doctors can quickly and clearly understand the information related to the inquiry points, which is convenient for doctors to make accurate and quick judgments and suggestions.

[0081] Step 205: Generate multiple candidate doctor dialogue information of the intelligent doctor for the current round of user dialogue information based on the current round of dialogue information and historical dialogue context through the consultation dialogue big model.

[0082] In this embodiment of the present application, if the key points of the consultation have not been cleared, that is, if no conversation has been obtained that meets the key point information acquisition requirements for all the key points of the consultation, the consultation dialogue macro model will generate multiple candidate doctor dialogue messages based on the current round of dialogue information and the historical dialogue context. These candidate dialogue messages are intended to guide the user to provide further information to meet the unmet key points of the consultation. By generating a variety of candidate dialogue messages, the system can flexibly respond to various user answers and reactions, improving the adaptability and interactivity of the consultation.

[0083] Step 206: Identify the key points of the candidate doctor conversation information, delete the candidate doctor conversation information containing the key points of the question, select one of the remaining candidate doctor conversation information as the next round of doctor conversation information to output, and use the next round of doctor conversation information and the user conversation information corresponding to the next round of doctor conversation information as the new current round conversation information, and return to the step of understanding the key points of the current round conversation information based on the key points of the question.

[0084] In an embodiment of the present application, the candidate doctor dialogue information is identified for the key points of the inquiry, thereby deleting the candidate dialogue information containing the key points of the inquiry that have been asked, and then the most suitable one is selected from the remaining candidate dialogue information as the next round of doctor dialogue information for output. At the same time, this new doctor dialogue information and its corresponding user dialogue information are used as the new dialogue information for this round, and the process of understanding the key points of the inquiry based on the key points of the inquiry is returned to continue the consultation process. By continuously iterating and optimizing the content of the conversation, the user's medical history information can be gradually and deeply obtained to ensure the comprehensiveness and depth of the consultation. At the same time, by deleting the dialogue information containing the key points of the inquiry that have been asked, redundancy and repetition are avoided, and the efficiency of the consultation is improved.

[0085] In summary, this intelligent medical history-based interview method, by combining a structured knowledge base, a large model for understanding key interview points, and a large model for interview dialogue, enables intelligent, efficient, and customized acquisition of patient medical history information. This method not only improves the accuracy and completeness of medical interviews but also enhances interactivity and communication efficiency between doctors and patients, providing strong support for the intelligent upgrade of medical services.

[0086] In the embodiment of the present application, optionally, step 205 includes:

[0087] Determine whether the current conversation round has reached the preset conversation round;

[0088] If the current conversation round does not reach the preset conversation round, the medical consultation conversation model is used to generate multiple candidate doctor conversation information for the current round of user conversation information based on the current round of conversation information and historical conversation context;

[0089] If the current conversation round has reached the preset conversation round, a conversation generation prompt word is constructed based on the medical inquiry key point information, the current round conversation information and the historical conversation context, and the conversation generation prompt word is input into the medical inquiry conversation model, so that the medical inquiry conversation model generates the next round of doctor conversation information containing any medical inquiry key point in the medical inquiry key point information based on the current round conversation information and the historical conversation context.

[0090] In the above embodiment, the current conversation turn between the intelligent doctor and the user is recorded. Before generating the next doctor conversation message, it is first determined whether the current conversation turn has reached the preset conversation turn limit. If the current conversation turn has not yet reached the preset conversation turn limit, the consultation conversation macromodel directly generates multiple candidate doctor conversation messages based on the current conversation information and historical conversation context. If the current conversation turn has reached or exceeded the preset conversation turn limit, a more refined generation strategy can be adopted. Specifically, a conversation generation prompt is first constructed based on the consultation key points information, the current conversation information, and the historical conversation context. This prompt contains the currently unmet consultation key points, the current conversation information, and the conversation context information, aiming to guide the consultation conversation macromodel to generate more targeted and accurate next doctor conversation messages. This prompt is then input into the consultation conversation macromodel, which, based on this information, generates the next doctor conversation message containing the unmet consultation key points. This ensures that the next doctor conversation message output by the macromodel both includes the uncovered consultation key points and maintains a coherent and unobtrusive consistency with the previous context. By setting preset conversation rounds, the embodiments of this application can effectively control the number and duration of conversations while ensuring the quality of the consultation, avoiding lengthy and repetitive consultation processes and improving consultation efficiency. After reaching the preset conversation rounds, by constructing conversation generation prompts and inputting them into the consultation dialogue model, the system can generate more targeted and accurate information for the next round of doctor conversations, ensuring the pertinence and depth of the consultation. Furthermore, through refined conversation generation strategies, the computing power of the consultation dialogue model can be more effectively utilized, reducing unnecessary computing overhead and optimizing resource utilization efficiency.

[0091] In an embodiment of the present application, optionally, the step 206 of selecting one of the remaining candidate doctor conversation information as the next round of doctor conversation information for output includes: if any of the remaining candidate doctor conversation information contains any of the consultation key points in the consultation key points information, then selecting one of the candidate doctor conversation information containing any of the consultation key points as the next round of doctor conversation information for output; if none of the remaining candidate doctor conversation information contains any of the consultation key points in the consultation key points information, then selecting one of the remaining candidate doctor conversation information as the next round of doctor conversation information for output.

[0092] In the above embodiment, when selecting the next round of doctor conversation information from the remaining candidate doctor conversation information, specifically, the remaining candidate doctor conversation information is first checked to determine whether it contains any of the key questions in the consultation key points information. This step is to ensure that the next round of doctor conversation information can continue the consultation process and guide the user to answer the unmet consultation key points. If there is at least one conversation message containing a consultation key point in the remaining candidate doctor conversation information, then one of the candidate conversation messages containing the consultation key points is selected as the next round of doctor conversation information for output. The selection criteria can be based on various factors, such as the clarity, relevance, and guidance of the conversation message, to ensure that the next round of conversation can effectively guide the user to answer the unmet consultation key points. If none of the remaining candidate doctor conversation information contains any of the key questions in the consultation key points information, then one of the candidate conversation messages not containing the consultation key points is selected as the next round of doctor conversation information for output. By preferentially selecting candidate conversation messages containing consultation key points as the next round of doctor conversation information, the present embodiment can ensure the continuity of the consultation process and avoid missing important consultation key points. Selecting one of the candidate dialogue messages containing the key points of the medical consultation for output can more effectively guide the user to answer the unmet medical consultation points, reduce unnecessary dialogue rounds, and improve medical consultation efficiency.

[0093] In an embodiment of the present application, optionally, before the medical consultation key points understanding large model understands the medical consultation key points of the current round of dialogue information based on the medical consultation key points information, the method further includes: obtaining a medical consultation dialogue information sample pre-labeled with a medical consultation key points label, wherein the medical consultation dialogue information sample includes at least one round of medical consultation dialogue, each round of medical consultation dialogue is labeled with a medical consultation key points label, the medical consultation key points label includes an empty label and a plurality of medical consultation key points identification labels, the empty label indicates that the corresponding round of medical consultation dialogue does not contain medical consultation key points or does not meet the key points information acquisition conditions, and the medical consultation key points identification label indicates the identification of the medical consultation key points contained in the corresponding round of medical consultation dialogue; fine-tuning the initial large model based on the medical consultation dialogue information sample to obtain the medical consultation key points understanding large model.

[0094] In the above embodiment, the specific training steps of the large model for understanding the key points of medical consultation are as follows: first, a series of medical consultation dialogue information samples pre-labeled with medical consultation key point labels are obtained. These samples include at least one round of medical consultation dialogue, and each round of medical consultation dialogue is labeled with medical consultation key point labels. Medical consultation key point labels are divided into empty labels and multiple medical consultation key point identification labels. An empty label indicates that the medical consultation dialogue in the corresponding round does not contain medical consultation key points or does not meet the conditions for obtaining key point information, that is, the round of dialogue does not provide any medical consultation key points that are helpful for diagnosis or the user's answer to the medical consultation key points does not comply with the rules. The medical consultation key point identification label indicates the identification of the specific medical consultation key points contained in the corresponding round of medical consultation dialogue. These identifications correspond to the various key points in the preset medical consultation key point information library. Then, these labeled medical consultation dialogue information samples are used to fine-tune the initial large model. Fine-tuning training is a method of further training for specific tasks or data sets based on a large-scale pre-trained model, aiming to improve the performance of the model on specific tasks. During the fine-tuning training process, the model learns to identify and understand the key information in the medical consultation dialogue, especially the information related to the key points of the medical consultation. Through continuous iterative training, the model can gradually improve its ability to recognize and accurately identify the key points of the medical consultation information. After fine-tuning training, a large model specifically for understanding the key points of the medical consultation is obtained, namely the large model for understanding the key points of the medical consultation. This model can deeply understand and analyze the medical consultation dialogue information based on the key points of the medical consultation information, determine whether the dialogue contains specific key points of the medical consultation, and evaluate whether these key points meet the conditions for obtaining the key point information. In the embodiment of the present application, by using labeled medical consultation dialogue information samples for fine-tuning training, the large model for understanding the key points of the medical consultation can more accurately identify and understand the key information in the medical consultation dialogue, especially the information related to the key points of the medical consultation. This helps to reduce the risk of misdiagnosis and missed diagnosis, and improve the accuracy and reliability of diagnosis. Fine-tuning training not only improves the performance of the model on specific tasks, but also enhances its generalization ability, enabling the large model to better adapt to medical consultation dialogues in different scenarios and situations, and improve its performance in complex situations. The application of a large-scale model for understanding key points in medical consultations can optimize the consultation process, enabling doctors to more efficiently obtain patient medical history information. By automatically identifying and understanding key points in medical consultations, the model can reduce unnecessary conversational turns, improve consultation efficiency, and more accurately understand the patient's medical history and needs, thereby providing more precise diagnosis and treatment recommendations, and enhancing patient satisfaction and trust.

[0095] By applying the technical solution of this embodiment, by introducing a large model for understanding the key points of medical consultation and a large model for medical consultation dialogue, and combining intelligent judgment of dialogue turns with careful selection of candidate dialogue information, the intelligent medical consultation method provided by the embodiment of this application significantly improves the accuracy and efficiency of medical consultation. The large model for understanding the key points of medical consultation can deeply understand and analyze the medical consultation dialogue information, accurately identify and evaluate the key points of medical consultation, and reduce the risk of misdiagnosis and missed diagnosis. At the same time, the large model for medical consultation dialogue can intelligently generate the next round of doctor dialogue information based on the historical dialogue context and the current round of dialogue information, guide the user to continue to provide information, and meet the unmet medical consultation points. In addition, by judging the dialogue turns, the system can flexibly adjust the dialogue generation strategy to ensure the continuity and efficiency of the medical consultation process. Finally, by carefully selecting candidate dialogue information, the system can provide more personalized and considerate medical consultation services, enhancing user satisfaction and trust. In summary, the embodiment of this application provides a more complete, efficient and reliable solution for the intelligent upgrade of medical services.

[0096] Further, as Figure 1 The specific implementation of the method, the embodiment of the present application provides an intelligent diagnosis device for the patient's medical history, such as Figure 3 As shown, the device includes:

[0097] a keyword recognition module configured to, in response to a request for the intelligent consultation service, output a first round of doctor conversation information of the intelligent doctor, obtain a first round of user conversation information for the first round of doctor conversation information, and perform medical history keyword recognition on the first round of user conversation information;

[0098] A medical question key point acquisition module is used to obtain, based on the identified medical history keywords, a plurality of medical question key points corresponding to the medical history keywords from a preset structured knowledge base as medical question key point information, wherein the preset structured knowledge base contains a plurality of medical question key points corresponding to each of the plurality of preset medical history keywords;

[0099] The intelligent consultation module is used to generate doctor dialogue information based on the first round of user dialogue information and the consultation key points information for output until the dialogue between the target user and the intelligent doctor covers all the consultation key points information.

[0100] Optionally, the medical consultation key points acquisition module is further used to:

[0101] Obtain the consultation category of the request initiation instruction;

[0102] A plurality of medical consultation points corresponding to the disease keywords under the consultation category are obtained from the preset structured knowledge base as the medical consultation point information.

[0103] Optionally, the intelligent medical consultation module is further used to:

[0104] The first round of doctor conversation information and the first round of user conversation information are used as current round conversation information. The current round conversation information is subjected to a key point understanding model based on the key point information to determine whether the current round conversation information satisfies a key point information obtaining condition for any key point in the key point information, wherein the key point information obtaining condition includes that the key point has been asked and an answer that complies with a rule has been obtained.

[0105] If the key point information obtaining condition of any medical consultation point is met, the satisfied medical consultation point will be removed from the medical consultation point information, and if the medical consultation point information is not cleared, the next round of doctor dialogue information will continue to be generated; otherwise, the next round of doctor dialogue information will stop being generated;

[0106] If the key information acquisition conditions of any of the medical consultation points are not met, the next round of doctor dialogue information is directly generated.

[0107] Optionally, the device further includes a model training module, configured to:

[0108] Obtain a sample of medical consultation dialogue information pre-labeled with a medical consultation key point label, wherein the medical consultation dialogue information sample includes at least one round of medical consultation dialogue, each round of medical consultation dialogue is labeled with a medical consultation key point label, and the medical consultation key point label includes an empty label and multiple medical consultation key point identification labels, wherein the empty label indicates that the corresponding round of medical consultation dialogue does not contain medical consultation key points or does not meet the key point information acquisition conditions, and the medical consultation key point identification label indicates the identification of the medical consultation key points contained in the corresponding round of medical consultation dialogue;

[0109] The initial large model is fine-tuned and trained based on the medical consultation dialogue information sample to obtain the large model for understanding the key points of the medical consultation.

[0110] Optionally, the intelligent medical consultation module is further used to:

[0111] The removed key points are recorded as key points asked;

[0112] Through the large model of medical consultation dialogue, based on the current round of dialogue information and historical dialogue context, the intelligent doctor generates multiple candidate doctor dialogue information for the current round of user dialogue information; the medical consultation key points of the candidate doctor dialogue information are identified, the candidate doctor dialogue information containing the medical consultation key points are deleted, and one of the remaining candidate doctor dialogue information is selected as the next round of doctor dialogue information for output, and the next round of doctor dialogue information and the user dialogue information corresponding to the next round of doctor dialogue information are used as the new current round of dialogue information, and the process returns to the step of understanding the medical consultation key points of the current round of dialogue information based on the medical consultation key points information.

[0113] Optionally, the intelligent medical consultation module is further used to:

[0114] Determine whether the current conversation round has reached the preset conversation round;

[0115] If the current conversation round does not reach the preset conversation round, the medical consultation conversation model is used to generate multiple candidate doctor conversation information for the current round of user conversation information based on the current round of conversation information and historical conversation context;

[0116] If the current conversation round has reached the preset conversation round, a conversation generation prompt word is constructed based on the medical inquiry key point information, the current round conversation information and the historical conversation context, and the conversation generation prompt word is input into the medical inquiry conversation model, so that the medical inquiry conversation model generates the next round of doctor conversation information containing any medical inquiry key point in the medical inquiry key point information based on the current round conversation information and the historical conversation context.

[0117] Optionally, the intelligent medical consultation module is further used to:

[0118] If any of the remaining candidate doctor dialogue information contains any of the medical inquiry points in the medical inquiry key points information, then one of the candidate doctor dialogue information containing any of the medical inquiry key points is selected as the next round of doctor dialogue information for output;

[0119] If the remaining candidate doctor dialogue information does not contain any of the medical inquiry points in the medical inquiry key points information, then one piece of the remaining candidate doctor dialogue information is selected as the next round of doctor dialogue information for output.

[0120] It should be noted that for other corresponding descriptions of the functional units involved in the intelligent medical inquiry device for patient medical history provided in the embodiment of the present application, please refer to Figures 1 to 2 The corresponding description in the method will not be repeated here.

[0121] The embodiment of the present application also provides a computer device, which can be specifically a personal computer, a server, a network device, etc. The computer device includes a bus, a processor, a memory and a communication interface, and may also include an input and output interface and a display device. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store location information. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the steps in each method embodiment are implemented.

[0122] Those skilled in the art will understand that the structure of the above-mentioned computer device is only a partial structure related to the solution of the present application and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components, or combine certain components, or have a different component arrangement.

[0123] In one embodiment, a computer-readable storage medium is provided. The computer-readable storage medium may be non-volatile or volatile, and stores a computer program thereon. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0124] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0125] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0126] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, and the like.

[0127] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0128] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. An intelligent diagnosis method for a patient's medical history, characterized in that: The method comprises: In response to a request for the intelligent consultation service, outputting first-round doctor conversation information of the intelligent doctor, obtaining first-round user conversation information for the first-round doctor conversation information, and performing medical history keyword recognition on the first-round user conversation information; Based on the identified medical history keywords, a plurality of medical inquiry key points corresponding to the medical history keywords are obtained from a preset structured knowledge base as medical inquiry key point information, wherein the preset structured knowledge base contains a plurality of medical inquiry key points corresponding to each of the plurality of preset medical history keywords; Based on the first round of user dialogue information and the key points of the consultation, doctor dialogue information is generated and outputted until the dialogue between the user and the intelligent doctor covers all key points of the consultation, including: The first round of doctor conversation information and the first round of user conversation information are used as current round conversation information. The current round conversation information is subjected to a key point understanding model based on the key point information to determine whether the current round conversation information satisfies a key point information obtaining condition for any key point in the key point information, wherein the key point information obtaining condition includes that the key point has been asked and an answer that complies with a rule has been obtained. If the key point information obtaining condition of any medical consultation point is met, the satisfied medical consultation point will be removed from the medical consultation point information, and if the medical consultation point information is not cleared, the next round of doctor dialogue information will continue to be generated; otherwise, the next round of doctor dialogue information will stop being generated; If the conditions for obtaining the key information of any of the key questions are not met, the next round of doctor dialogue information will be directly generated; After removing the satisfied medical inquiry points from the medical inquiry points information, the method further includes: The removed key points are recorded as key points asked; Generate the next round of doctor dialogue information, including: Generate multiple candidate doctor dialogue messages for the current round of user dialogue information based on the current round of dialogue information and historical dialogue context through the consultation dialogue big model; Identify the key points of the candidate doctor conversation information, delete the candidate doctor conversation information containing the key points of the question, select one of the remaining candidate doctor conversation information as the next round of doctor conversation information for output, and use the next round of doctor conversation information and the user conversation information corresponding to the next round of doctor conversation information as the new current round conversation information, and return to the step of understanding the key points of the current round conversation information based on the key points of the question.

2. The method according to claim 1, characterized in that The method of obtaining, based on the identified medical history keywords, a plurality of medical inquiry key points corresponding to the medical history keywords in a preset structured knowledge base as medical inquiry key point information includes: Obtain the consultation category of the request initiation instruction; A plurality of medical inquiry points corresponding to the medical history keywords under the consultation category are obtained from the preset structured knowledge base as the medical inquiry point information.

3. The method according to claim 1, characterized in that Before the large model for understanding the key points of the medical inquiry is used to understand the key points of the medical inquiry information of the current round of dialogue information based on the key points of the medical inquiry information, the method further includes: Obtain a sample of medical consultation dialogue information pre-labeled with a medical consultation key point label, wherein the medical consultation dialogue information sample includes at least one round of medical consultation dialogue, each round of medical consultation dialogue is labeled with a medical consultation key point label, and the medical consultation key point label includes an empty label and multiple medical consultation key point identification labels, wherein the empty label indicates that the corresponding round of medical consultation dialogue does not contain medical consultation key points or does not meet the key point information acquisition conditions, and the medical consultation key point identification label indicates the identification of the medical consultation key points contained in the corresponding round of medical consultation dialogue; The initial large model is fine-tuned and trained based on the medical consultation dialogue information sample to obtain the large model for understanding the key points of the medical consultation.

4. The method according to claim 1, wherein The large model of the medical consultation dialogue generates multiple candidate doctor dialogue information for the current round of user dialogue information based on the current round of dialogue information and the historical dialogue context, including: Determine whether the current conversation round has reached the preset conversation round; If the current conversation round does not reach the preset conversation round, the medical consultation conversation model is used to generate multiple candidate doctor conversation information for the current round of user conversation information based on the current round of conversation information and historical conversation context; If the current conversation round has reached the preset conversation round, a conversation generation prompt word is constructed based on the medical inquiry key point information, the current round conversation information and the historical conversation context, and the conversation generation prompt word is input into the medical inquiry conversation model, so that the medical inquiry conversation model generates the next round of doctor conversation information containing any medical inquiry key point in the medical inquiry key point information based on the current round conversation information and the historical conversation context.

5. The method according to claim 1, wherein The step of selecting one piece of the remaining candidate doctor dialogue information as the next round of doctor dialogue information for output includes: If any of the remaining candidate doctor dialogue information contains any of the medical inquiry points in the medical inquiry key points information, then one of the candidate doctor dialogue information containing any of the medical inquiry key points is selected as the next round of doctor dialogue information for output; If the remaining candidate doctor dialogue information does not contain any of the medical inquiry key points in the medical inquiry key points information, then one piece of the remaining candidate doctor dialogue information is selected as the next round of doctor dialogue information for output.

6. An intelligent medical consultation device for a patient's medical history, characterized in that: The device comprises: a keyword recognition module configured to, in response to a request for the intelligent consultation service, output a first round of doctor conversation information of the intelligent doctor, obtain a first round of user conversation information for the first round of doctor conversation information, and perform medical history keyword recognition on the first round of user conversation information; A medical question key point acquisition module is used to obtain, based on the identified medical history keywords, a plurality of medical question key points corresponding to the medical history keywords from a preset structured knowledge base as medical question key point information, wherein the preset structured knowledge base contains a plurality of medical question key points corresponding to each of the plurality of preset medical history keywords; An intelligent consultation module, configured to generate and output doctor dialogue information based on the first round of user dialogue information and the consultation key points, until the dialogue between the user and the intelligent doctor covers all the consultation key points; The intelligent medical consultation module is further used to: The first round of doctor conversation information and the first round of user conversation information are used as current round conversation information. The current round conversation information is subjected to a key point understanding model based on the key point information to determine whether the current round conversation information satisfies a key point information obtaining condition for any key point in the key point information, wherein the key point information obtaining condition includes that the key point has been asked and an answer that complies with a rule has been obtained. If the key point information obtaining condition of any medical consultation point is met, the satisfied medical consultation point will be removed from the medical consultation point information, and if the medical consultation point information is not cleared, the next round of doctor dialogue information will continue to be generated; otherwise, the next round of doctor dialogue information will stop being generated; If the conditions for obtaining the key information of any of the key questions are not met, the next round of doctor dialogue information will be directly generated; The intelligent medical consultation module is further used to: The removed key points are recorded as key points asked; Through the large model of medical consultation dialogue, based on the current round of dialogue information and historical dialogue context, the intelligent doctor generates multiple candidate doctor dialogue information for the current round of user dialogue information; the medical consultation key points of the candidate doctor dialogue information are identified, the candidate doctor dialogue information containing the medical consultation key points are deleted, and one of the remaining candidate doctor dialogue information is selected as the next round of doctor dialogue information for output, and the next round of doctor dialogue information and the user dialogue information corresponding to the next round of doctor dialogue information are used as the new current round of dialogue information, and the process returns to the step of understanding the medical consultation key points of the current round of dialogue information based on the medical consultation key points information.

7. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

8. A computer device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.

Images