A guided medical consultation method, system, electronic device and storage medium based on a large language model

Abstract

The present invention provides a guided medical consultation method, system, electronic device and storage medium based on a large language model. By introducing the large language model, the patient's self-reported symptoms can be quickly understood and processed. Compared with traditional manual medical consultation, the consultation time is greatly shortened and the diagnosis and treatment efficiency is improved. By utilizing the knowledge graph of the relationship between traditional Chinese medicine symptoms, the patient's symptoms can be accurately matched with the relevant symptoms in traditional Chinese medicine theory. Through big data analysis and model learning, the accuracy of the medical consultation can be gradually optimized, and more accurate diagnostic suggestions can be provided to the patient. According to the patient's initial feedback, subsequent symptom inquiries can be intelligently raised, simulating the meticulous medical consultation process of traditional Chinese medicine practitioners, ensuring the comprehensiveness, depth and interpretability of the medical consultation.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical technology, and in particular to a guided medical consultation method, system, electronic device and storage medium based on a large language model. Background Art

[0002] Among the four examinations in traditional Chinese medicine, medical interview occupies an important position and has been highly valued by doctors of all generations. Medical interview in traditional Chinese medicine is a diagnostic method in which doctors ask patients or accompanying persons questions to understand the occurrence, development, treatment process, current symptoms and other relevant conditions of the disease. By communicating with the patient, the doctor can obtain the patient's chief symptoms and gradually make a diagnosis during the interview until information supporting diagnosis and treatment is obtained. Effective medical interview can collect symptoms that patients are unwilling to talk about, do not know how to express or think are unimportant, which is of great significance for the diagnosis and treatment of diseases in traditional Chinese medicine.

[0003] However, traditional Chinese medicine consultation relies on the doctor's rich clinical experience, is highly subjective, and lacks unified standards. Therefore, research on the objectivity of consultation is of great significance to improving the efficiency and efficacy of diagnosis and treatment, as well as the inheritance and innovation of Chinese medicine clinical experience.

[0004] With the development of artificial intelligence (AI) technology, knowledge graphs and large language models are gradually being applied to Traditional Chinese Medicine (TCM) diagnosis, enabling more people to enjoy intelligent TCM consultation services. However, existing methods often rely on hard matching with synonym databases. If a patient's description doesn't match a synonym database, it can't be converted into vocabulary that the model can understand, making the consultation impossible. Existing methods also lack TCM syndrome differentiation logic in the questioning and thinking process. Patient symptom terminology is vague, making it difficult for machines to understand colloquial symptoms, resulting in low human-computer interaction accuracy. Furthermore, these methods are often oriented towards a single disease, limiting their scope of application. Summary of the Invention

[0005] In order to solve the problems existing in the prior art, the present invention provides a guided medical consultation method, system, electronic device and storage medium based on a large language model, which can accurately understand the user's natural language expression, dynamically adjust the inquiry content, gradually obtain the user's detailed needs, ensure the accuracy and interpretability of the consultation results, and significantly improve the intelligence level of traditional Chinese medicine consultation and user experience.

[0006] In a first aspect, the present invention provides a guided medical consultation method based on a large language model, comprising the following steps:

[0007] Step 1: Obtain the symptom content entered by the user;

[0008] Step 2: Using the retrieval enhancement generation (RAG) technology, n symptom terms related to the user's symptom content are extracted from the preset symptom term knowledge base as n associated symptom labels corresponding to the user's symptom content;

[0009] Step 3: Use the large language model to map the user's symptom content and the corresponding n associated symptom labels, obtain the mapped symptom labels corresponding to the user's symptom content, and add them to the combined symptom label set;

[0010] Step 4: Based on the combined symptom label set and the TCM disease relationship knowledge graph, determine the symptom label with the highest priority relevance to the combined symptom label set as the current query symptom label, and determine whether the preset query termination condition is met;

[0011] Step 5: If the preset inquiry termination condition is met, the corresponding diagnosis result is output based on the combined symptom label set; otherwise, the corresponding symptom inquiry question is generated based on the current inquiry symptom label and the inquiry question is output to the user. Steps 1 to 4 are executed repeatedly until the preset inquiry termination condition is met, and the corresponding diagnosis result is output based on the current combined symptom label set.

[0012] In a second aspect, the present invention provides a guided medical consultation system based on a large language model, comprising:

[0013] The acquisition module is used to obtain the symptom content input by the user;

[0014] A label extraction module is used to extract n symptom terms related to the user's symptom content from a preset symptom term knowledge base using the retrieval enhancement generation (RAG) technology as n associated symptom labels corresponding to the user's symptom content;

[0015] The label mapping module is used to map the user's symptom content and the corresponding n associated symptom labels using a large language model, obtain the mapped symptom label corresponding to the user's symptom content, and add it to the combined symptom label set;

[0016] A label relevance calculation module is used to determine the symptom label with the highest relevance priority to the combined symptom label set as the current inquiry symptom label based on the combined symptom label set and the TCM disease relationship knowledge graph;

[0017] A termination judgment module is used to judge whether the preset inquiry termination conditions are met;

[0018] The inquiry question output module is used to generate corresponding symptom inquiry questions based on the current inquiry symptom label and output the inquiry questions to the user;

[0019] The diagnosis result output module is used to output the corresponding diagnosis result based on the current combined symptom label set.

[0020] In a third aspect, the present invention provides an electronic device comprising: at least one processor and a memory;

[0021] The memory stores computer-executable instructions;

[0022] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the guided diagnosis method based on the large language model as described in the first aspect.

[0023] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions. When a processor executes the computer-executable instructions, the guided medical consultation method based on the large language model as described in the first aspect is implemented.

[0024] After adopting the above technical solution, the present invention has the following beneficial effects:

[0025] 1. Improve diagnostic efficiency: By introducing a large language model, it can quickly understand and process patients' self-reported symptoms. Compared with traditional manual consultation, it greatly shortens the consultation time and improves diagnosis and treatment efficiency;

[0026] 2. Improved accuracy: Utilizing the TCM disease relationship knowledge graph, we can accurately match patient symptoms with relevant symptoms in TCM theory. Through big data analysis and model learning, we can gradually optimize the accuracy of consultations and provide patients with more accurate diagnostic recommendations.

[0027] 3. Intelligent medical consultation: Based on the patient's initial feedback, the system can intelligently ask follow-up symptom inquiries, simulating the meticulous medical consultation process of a traditional Chinese medicine practitioner, ensuring the comprehensiveness, depth and explainability of the medical consultation. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] 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 only 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.

[0029] Figure 1 A flowchart of a guided medical consultation method based on a large language model provided in the first embodiment of the present disclosure;

[0030] Figure 2 A schematic diagram of the structure of a guided medical consultation system based on a large language model provided in the second embodiment of the present disclosure;

[0031] Figure 3 This is a structural diagram of an electronic device provided in Example 3 of the present disclosure. DETAILED DESCRIPTION

[0032] The present disclosure will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the present disclosure and are not intended to limit the present disclosure. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions relevant to the present disclosure, not all structures.

[0033] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the steps as sequential processes, many of the steps can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the steps can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the accompanying drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0034] In the following embodiments, each embodiment provides optional features and examples at the same time. The various features recorded in the embodiments can be combined to form multiple optional solutions. Each numbered embodiment should not be regarded as just one technical solution.

[0035] Example 1

[0036] The present disclosure provides a guided medical consultation method based on a large language model, specifically including:

[0037] 1. Obtain the symptom content entered by the user for the first time;

[0038] In this embodiment, the user's initial symptom content is first received. The user can provide their symptom description through text input, voice input or other forms of input. For example, the user can enter "I have been having back pain frequently recently and I don't know why" in the text box.

[0039] 2. Using the retrieval enhancement generation (RAG) technology, n symptom terms related to the user's symptom content are extracted from the preset symptom term knowledge base as n associated symptom labels corresponding to the user's symptom content;

[0040] In this embodiment, the retrieval enhanced generation RAG technology combines the advantages of information retrieval and generation models to improve the accuracy and relevance of label retrieval. For example, for the user input "I have low back pain recently", related symptom labels such as "low back pain", "weak waist and knees", "low back pain due to falls and sprains", and "back pain" may be retrieved.

[0041] 3. Use the large language model to map the user's symptom content and the corresponding n associated symptom labels, obtain the mapped symptom labels corresponding to the user's symptom content, and add them to the combined symptom label set;

[0042] In this embodiment, a large language model is used to map the user's symptom content to the label obtained in the previous step. Through this process, the user's description can be understood more accurately and converted into a format that conforms to the symptom term knowledge base label. For example, the user input "I recently had low back pain" will be mapped to the "low back pain" label. During the initial inquiry, there is no label in the combined symptom label set, and the "low back pain" label is added to the combined symptom label set.

[0043] 4. Based on the combined symptom label set and the TCM disease relationship knowledge graph, determine the symptom label with the highest priority in correlation with the combined symptom label set as the current inquiry symptom label;

[0044] In this embodiment, based on the TCM disease relationship knowledge graph, the relevance priority of each label is calculated. The TCM disease relationship knowledge graph includes the relationship between TCM syndromes and symptoms, the relationship between symptoms and symptoms, and the relationship between TCM syndromes and disease locations. The calculation method of symptom label priority is not limited. The symptom label that is most relevant to the combined symptom label set can be solved by calculation to proceed to the next inquiry. For example, it is calculated that the "weak waist and knees" label has the highest correlation with the "low back pain" label.

[0045] 5. Generate corresponding symptom inquiry questions based on the current inquiry symptom label and output the inquiry questions to the user;

[0046] In this embodiment, the next inquiry question is generated based on the most relevant symptom tag. These inquiry questions are intended to guide the user to provide further specific information, thereby helping the system better understand the user's needs or problems. The inquiry questions can be output in the form of text or voice, and require the user to answer in a certain way. For example, the system asks "Do you have any soreness in your waist and knees?"

[0047] 6. Get the symptom content entered by the user for the second time

[0048] In this embodiment, the user makes a second input based on the inquiry question provided by the system. The user's answer will further enrich the system's understanding of the question and provide more information for subsequent analysis. For example, the user answers "I usually feel sore in the waist and knees."

[0049] 7. Multi-symptom label correlation calculation

[0050] In this embodiment, the mapped symptom labels generated by the user's second input are combined with the previously obtained mapped symptom labels, and the symptom label relevance priorities are recalculated. The priorities of the symptom labels are dynamically adjusted to ensure that each query is based on the latest correlation analysis results. For example, in the second calculation, the "swelling" label has the highest correlation with the "weak waist and knees" label and the "low back pain" label.

[0051] 8. Circular Inquiry

[0052] In this embodiment, based on each answer of the user, the system repeats the correlation calculation and inquiry steps until the preset inquiry termination condition is met. The inquiry termination condition may be that a certain confidence threshold is reached, sufficient information is obtained, or the user explicitly indicates that no further inquiry is required.

[0053] 9. Output corresponding diagnostic results based on the combined symptom label set

[0054] In this embodiment, after the inquiry is completed, the final combined symptom label set is obtained based on the results of multiple rounds of inquiries, and the final diagnostic suggestions or results are output based on the final combined symptom label set. These results can be detailed answers to user questions, solution suggestions, or other forms of output. For example, the final combined symptom label set includes an "edema" label, a "weak waist and knees" label, and a "low back pain" label, and the output is "Based on your description and the results of multiple rounds of inquiries, it is recommended that you seek medical attention to check whether there is a problem of kidney yin deficiency, and consider relevant treatment."

[0055] Through the above steps, the present invention realizes an efficient guided interactive inquiry method, which can accurately understand the user's natural language expression, dynamically adjust the inquiry content, gradually obtain the user's detailed needs, ensure the accuracy and interpretability of the inquiry results, and significantly improve the intelligence level of the inquiry and user experience.

[0056] Example 2

[0057] refer to Figure 2 The present disclosure provides a guided medical consultation system based on a large language model, including:

[0058] An acquisition module 21 is used to acquire the symptom content input by the user;

[0059] The tag extraction module 22 is used to extract n symptom terms related to the user's symptom content from a preset symptom term knowledge base using the retrieval enhancement generation (RAG) technology as n associated symptom tags corresponding to the user's symptom content;

[0060] The label mapping module 23 is used to map the user's symptom content and the corresponding n associated symptom labels using a large language model, obtain the mapped symptom label corresponding to the user's symptom content, and add it to the combined symptom label set;

[0061] A label relevance calculation module 24 is configured to determine, based on the combined symptom label set and the TCM disease relationship knowledge graph, the symptom label with the highest relevance priority to the combined symptom label set as the current inquiry symptom label;

[0062] The termination judgment module 25 is used to judge whether the preset inquiry termination condition is met;

[0063] The inquiry question output module 26 is used to generate a corresponding symptom inquiry question based on the current inquiry symptom tag and output the inquiry question to the user;

[0064] The diagnosis result output module 27 is used to output the corresponding diagnosis result based on the current combined symptom label set.

[0065] Example 3

[0066] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure is shown in FIG. Figure 3 As shown, the electronic device 30 of this embodiment may include: a memory 31 and a processor 32.

[0067] Memory 31, for storing computer programs (such as application programs and functional modules for implementing the aforementioned guided medical consultation method based on a large language model), computer instructions, etc.;

[0068] The aforementioned computer programs, computer instructions, etc. may be partitioned and stored in one or more memories 31 . Furthermore, the aforementioned computer programs, computer instructions, data, etc. may be called by the processor 32 .

[0069] The processor 32 is configured to execute the computer program stored in the memory 31 to implement the various steps in the method involved in the above embodiment.

[0070] For details, please refer to the relevant description in the previous method embodiment.

[0071] The memory 31 and the processor 32 may be independent structures or integrated structures. When the memory 31 and the processor 32 are independent structures, the memory 31 and the processor 32 may be coupled to each other via a bus 33 .

[0072] An electronic device of this embodiment can perform Figure 1 The technical solution in the method shown, its specific implementation process and technical principles can be found in Figure 1 The relevant descriptions in the method shown will not be repeated here.

[0073] In addition, an embodiment of the present application further provides a computer-readable storage medium, in which computer-executable instructions are stored. When at least one processor of a user device executes the computer-executable instructions, the user device executes the various possible methods described above.

[0074] Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transfer of computer programs from one location to another. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer. An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and storage medium can be located in an ASIC. Alternatively, the ASIC can be located in a user device. Of course, the processor and storage medium can also exist as discrete components in a communication device.

[0075] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0076] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A guided medical consultation method based on a large language model, characterized in that: The following steps are involved: Step 1: Obtain the symptom content entered by the user; Step 2: Using the retrieval enhancement generation (RAG) technology, n symptom terms related to the user's symptom content are extracted from the preset symptom term knowledge base as n associated symptom labels corresponding to the user's symptom content; Step 3: Use the large language model to map the user's symptom content and the corresponding n associated symptom labels, obtain the mapped symptom labels corresponding to the user's symptom content, and add them to the combined symptom label set; Step 4: Based on the combined symptom label set and the TCM disease relationship knowledge graph, determine the symptom label with the highest priority relevance to the combined symptom label set as the current query symptom label, and determine whether the preset query termination condition is met; Step 5: If the preset inquiry termination condition is met, the corresponding diagnosis result is output based on the combined symptom label set; otherwise, the corresponding symptom inquiry question is generated based on the current inquiry symptom label and the inquiry question is output to the user. Steps 1 to 4 are executed repeatedly until the preset inquiry termination condition is met, and the corresponding diagnosis result is output based on the current combined symptom label set.

2. A guided medical consultation system based on a large language model, characterized in that: include: The acquisition module is used to obtain the symptom content input by the user; A label extraction module is used to extract n symptom terms related to the user's symptom content from a preset symptom term knowledge base using the retrieval enhancement generation (RAG) technology as n associated symptom labels corresponding to the user's symptom content; The label mapping module is used to map the user's symptom content and the corresponding n associated symptom labels using a large language model, obtain the mapped symptom label corresponding to the user's symptom content, and add it to the combined symptom label set; A label relevance calculation module is used to determine the symptom label with the highest relevance priority to the combined symptom label set as the current inquiry symptom label based on the combined symptom label set and the TCM disease relationship knowledge graph; A termination judgment module is used to judge whether the preset inquiry termination conditions are met; The inquiry question output module is used to generate corresponding symptom inquiry questions based on the current inquiry symptom label and output the inquiry questions to the user; The diagnosis result output module is used to output the corresponding diagnosis result based on the current combined symptom label set.

3. An electronic device, characterized in that: include: at least one processor and memory; The memory stores computer-executable instructions; The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the guided diagnosis method based on a large language model according to claim 1.

4. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions. When the processor executes the computer-executable instructions, the guided diagnosis method based on the large language model as claimed in claim 1 is implemented.

Images