Medical evidence-based method and device and computer equipment
By introducing authority labels into the large language model, the problem of inaccurate medical evidence-based output responses is solved, improving the authority of the information and the user experience.
Patent Information
- Application Number
- CN202511156783.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-18
- Publication Date
- 2025-12-16
AI Technical Summary
Existing large language models lack authority and accuracy in providing output responses in medical evidence-based medicine, resulting in a poor user experience.
An authority label is introduced, which indicates the authority of evidence through evidence number and authority label. Response information is generated and displayed, including information such as the research type, source, journal level, and time of the evidence.
This improves the authority and accuracy of large language models in providing evidence-based medical responses, allowing users to quickly identify the authority of responses and enhancing the user experience.
Smart Images

Figure CN121148644A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of evidence-based medicine, and more particularly to a medical evidence-based method, apparatus, and computer device. Background Technology
[0002] Evidence-Based Medicine (EBM) in medicine refers to a medical practice model that combines the best current research evidence, the professional skills of clinicians, and the personal values and needs of patients to guide clinical decision-making. Its core principle is to replace experiential intuition with scientific evidence, ensuring the objectivity and effectiveness of medical decisions.
[0003] Due to the illusion problem of existing large language models and the limitations of search technology, the authority and accuracy of large language models in providing evidence-based medical responses are not high, resulting in a poor user experience. Summary of the Invention
[0004] In view of this, embodiments of the present invention provide a medical evidence-based method, apparatus, and computer device. By introducing authority labels, the authority and accuracy of the large language model's output response information for medical evidence-based purposes can be improved. Users can quickly identify the authority of the response information based on the authority labels, thereby enhancing the user experience.
[0005] In a first aspect, embodiments of the present invention provide a medical evidence-based method applied to a human-computer dialogue application based on a large language model, the method comprising:
[0006] Receive first information related to medical evidence-based practice input by the user in the human-computer dialogue interface of the human-computer dialogue application;
[0007] Generate and display a response message for the first information, the response message including:
[0008] The first cited content and the first evidence label corresponding to the first cited content;
[0009] The first evidence label corresponding to the first cited content includes an evidence number label and at least one authority label, wherein the authority label is used to indicate the authority of the evidence corresponding to the cited content.
[0010] Optionally, the response information further includes: second cited content and a second evidence tag corresponding to the second cited content, wherein the second evidence tag only includes the evidence sequence number tag, and the authority of the evidence corresponding to the first evidence tag that includes at least one authority tag is higher than the authority of the evidence corresponding to the second evidence tag that only includes the evidence sequence number tag.
[0011] Optionally, the response information further includes: second cited content and a second evidence tag corresponding to the second cited content, wherein the second evidence tag only includes the evidence sequence number tag, and the sequence number of the evidence sequence number tag in the first evidence tag is less than the sequence number of the evidence sequence number tag in the second evidence tag.
[0012] Optionally, the authority label includes at least one of the following: research type label, source label, and journal level label.
[0013] Optionally, the authority label may also include a time label for the evidence.
[0014] Optionally, the method further includes:
[0015] The evidence is ordered according to the priority of the pre-set authority label to obtain the serial number of the evidence corresponding to the multiple cited contents;
[0016] The evidence number label displays the number of the evidence corresponding to the cited content.
[0017] Optionally, generating the response information for the first information includes:
[0018] Based on the first information and the knowledge base, the response information is generated using the large language model; the knowledge base contains multiple different pieces of evidence and the authority labels corresponding to the evidence.
[0019] Optionally, the evidence label may further include an evidence format label.
[0020] Optionally, the format of the evidence format label and the evidence serial number label is a superscript.
[0021] Optionally, the evidence format label is a PDF label, and the method further includes:
[0022] In response to the user's click operation on the evidence tag with the PDF tag, a PDF preview window pops up and displays the page of the evidence corresponding to the evidence tag in the PDF preview window, and the cited content in the PDF preview window is highlighted.
[0023] Optionally, the method further includes:
[0024] In response to the user's selection of a portion of the evidence in the PDF preview window, at least one related control is displayed, including a translation control, a copy control, or a search control;
[0025] In response to the user's operation on the relevant control, the operation corresponding to the relevant control is performed on the part of the content.
[0026] Optionally, the response information further includes: an evidence list control; the method further includes:
[0027] In response to the user's click operation on the evidence list control, an evidence list pops up; the evidence list includes: evidence cards corresponding to the cited content, and the evidence cards contain the authority label of the evidence corresponding to the cited content.
[0028] Optionally, the cited content comes from a journal, and the evidence card also includes an impact factor (IF) label, which indicates the impact value of the journal.
[0029] Optionally, the evidence card includes multiple authority labels for the evidence, and the multiple authority labels are ordered according to their priority.
[0030] Optionally, the evidence card further includes: a reason for including the evidence corresponding to the cited content, wherein the reason for inclusion is generated by the large language model through analysis and summary of the first information and the evidence.
[0031] Optionally, the evidence cards in the evidence list are sorted according to their authority.
[0032] Optionally, the response information further includes: a first evidence control, the first evidence control containing the serial number, name and source information of a first piece of evidence, the first piece of evidence being the first piece of evidence in the evidence list;
[0033] The method further includes:
[0034] In response to the user's operation on the first evidence control, the page of the first evidence is opened.
[0035] Optionally, the response information further includes: a mind chain expansion control; the mind chain is displayed in a waterfall layout before the response information is displayed and collapses when the response information is started to be displayed; the expansion control is used to trigger the expansion of the mind chain in response to user operation.
[0036] Optionally, the evidence label is displayed in a highlighted manner.
[0037] On the other hand, embodiments of the present invention provide a medical evidence-based device, the device comprising:
[0038] The receiving module is used to receive the first information related to medical evidence input by the user in the human-computer dialogue interface; the human-computer dialogue interface is the interface in the human-computer dialogue application based on the large language model.
[0039] The processing module is used to generate the response information for the first information;
[0040] A display module is used to display the response information, which includes:
[0041] The first cited content and the first evidence label corresponding to the first cited content;
[0042] The first evidence label corresponding to the first cited content includes an evidence number label and at least one authority label, wherein the authority label is used to indicate the authority of the evidence corresponding to the cited content.
[0043] On the other hand, embodiments of the present invention provide a storage medium including a stored program, wherein the program controls the device where the storage medium is located to execute the above-described method when it is running.
[0044] On the other hand, embodiments of the present invention provide a computer device including a memory and a processor. The memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions, wherein the program instructions, when loaded and executed by the processor, implement the steps of the above-described method.
[0045] In the technical solutions of the medical evidence-based method, apparatus, and computer device provided in this invention, the medical evidence-based method is applied to a human-computer dialogue application based on a large language model. The method includes: receiving first information related to medical evidence-based practice input by a user in the human-computer dialogue interface of the application; generating and displaying response information for the first information, the response information including: first cited content and a first evidence tag corresponding to the first cited content; wherein the first evidence tag corresponding to the first cited content includes an evidence sequence number tag and at least one authority tag, the authority tag indicating the authority of the evidence corresponding to the cited content. By introducing an authority tag, this invention can improve the authority and accuracy of the response information output by the large language model for medical evidence-based practice. Users can quickly identify the authority of the response information based on the authority tag, enhancing the user experience. Attached Figure Description
[0046] Figure 1 A flowchart of a medical evidence-based method provided in an embodiment of the present invention;
[0047] Figure 2 This is a schematic diagram of the process for generating the response information of the first information in an embodiment of the present invention;
[0048] Figure 3 This is a schematic diagram of the interface displaying the reply information of the first information in an embodiment of the present invention;
[0049] Figure 4 A flowchart of yet another medical evidence-based method provided in an embodiment of the present invention;
[0050] Figure 5 A flowchart of yet another medical evidence-based method provided in an embodiment of the present invention;
[0051] Figure 6 This is a schematic diagram of the pop-up evidence list in an embodiment of the present invention;
[0052] Figure 7 A flowchart of yet another medical evidence-based method provided in an embodiment of the present invention;
[0053] Figure 8 This is a schematic diagram of the structure of a medical evidence-based device provided in an embodiment of the present invention;
[0054] Figure 9 This is a schematic diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation
[0055] To better understand the technical solution of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
[0056] It should be understood that the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.
[0057] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.
[0058] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.
[0059] Figure 1 A flowchart of a medical evidence-based method provided in this embodiment of the invention, applied to a human-computer dialogue application based on a large language model, such as... Figure 1 As shown, the method includes:
[0060] Step 101: Receive the first information related to medical evidence-based practice entered by the user in the human-computer dialogue interface of the human-computer dialogue application.
[0061] In this embodiment of the invention, each step can be performed by a computer device.
[0062] For example, Figure 2 This is a schematic diagram of the process for generating the response information of the first information in an embodiment of the present invention, such as... Figure 2 As shown, the system first performs a demand assessment on the user's initial input, the prompt, to determine if it meets the conditions for triggering evidence-based research. If the conditions are met, the evidence-based research process is initiated. For example, the conditions for triggering evidence-based research include: the initial input is relevant to medical evidence. Then, the evidence-based research process is executed. The demand assessment function is simple and can utilize a small-sized large language model, reducing costs, such as a 7B model (e.g., LLaMA-7B or DeepSeek-7B). In a 7B model, the "B" stands for Billion, meaning a model with 7 billion parameters.
[0063] For example, "related to medical evidence-based practice" refers to any one or any combination of aspects such as clinical diagnosis and treatment support, medical teaching, medical research, improving medical efficiency, and basic medical knowledge. For instance... Figure 2 As shown, the evidence-based process includes intent understanding, evidence retrieval, and generating response results. In the intent understanding process, a thought chain is first generated based on initial information. A set of questions is then derived from the thought chain. Natural language subqueries are generated from the questions in the question set, and these subqueries are searched in both Chinese and English to obtain multiple queries. Generating the thought chain requires strong reasoning capabilities, so a large language model can be chosen to improve the accuracy of reasoning, such as DeepSeek's R1'-671B, where the 671B model represents a model with 671 billion parameters. The functionality of generating the question set from the thought chain and generating multiple queries from the question set is of medium level, and a medium-sized large language model can be used to balance functional requirements and cost, such as the 72B model (e.g., LLaMA-72B), which represents a model with 72 billion parameters.
[0064] Then, as Figure 2As shown, in the evidence retrieval process, multiple queries are retrieved based on a pre-built knowledge base to recall a set of evidence. Then, the evidence set is refined based on the relevance of the evidence to the query and the authority of the evidence, resulting in multiple evidence documents. These documents are then analyzed to arrive at a conclusion. The retrieval, recall, and refinement functions are simple and can utilize small-sized large language models to reduce costs, such as the 7B model (e.g., LLaMA-7B or DeepSeek-7B), where a 7B model represents a model with 7 billion parameters. The evidence analysis function is of a medium level and can utilize medium-sized large language models to balance functional requirements and cost, such as the 72B model (e.g., LLaMA-72B), where a 72B model represents a model with 72 billion parameters.
[0065] Step 102: Generate and display the response information for the first information. The response information includes: the first cited content and the first evidence label corresponding to the first cited content; wherein, the first evidence label corresponding to the first cited content includes an evidence serial number label and at least one authority label, and the authority label is used to indicate the authority of the evidence corresponding to the cited content.
[0066] By introducing authority labels, this invention can improve the authority and accuracy of large language models in outputting medical evidence-based responses. Users can quickly identify the authority of the responses based on the authority labels, thus enhancing the user experience.
[0067] For example, such as Figure 2 As shown, in the process of generating response information, the first information answer is generated based on the conclusion. The answer includes the response information and multiple pieces of evidence. Generating the answer requires strong reasoning ability, so a large language model can be selected to improve the accuracy of the reasoning, such as DeepSeek's R1'-671B. The 671B model represents a model with 671 billion parameters.
[0068] Compared to ordinary search scenarios, the medical evidence-based approach in this invention is mainly aimed at doctors, medical students, and medical research groups for clinical decision-making or scientific research. It has very high requirements for the authority of the evidence provided by the evidence-based approach. Therefore, it is necessary to stratify the searched reference materials according to their authority and credibility levels, so as to directly find authoritative evidence content related to the primary information and provide it to the user.
[0069] Optionally, the response information may also include: second cited content and a second evidence label corresponding to the second cited content, wherein the second evidence label only includes an evidence number label, and the evidence corresponding to the first evidence label including at least one authority label has a higher authority than the evidence corresponding to the second evidence label which only includes an evidence number label.
[0070] For example, medical guidelines have higher authority than drug instructions. When the evidence corresponding to the cited content is a medical guideline, an authority label will be displayed in the evidence label of that cited content; for example, the authority label will include the text "Guideline". When the evidence corresponding to the cited content is a drug instruction, an authority label will not be displayed in the evidence label of that cited content.
[0071] In this embodiment of the invention, since the response information for medical evidence-based practice includes an authority label for the evidence, the authority of the evidence corresponding to the evidence label including the authority label is higher than that of the evidence corresponding to the evidence label containing only the evidence number label. Therefore, the authority and accuracy of the response information for medical evidence-based practice can be improved, meeting the requirements of medical users and enhancing the user experience.
[0072] Optionally, the response information may also include: the second cited content and a second evidence label corresponding to the second cited content. The second evidence label only includes an evidence sequence number label, wherein the sequence number of the evidence in the first evidence label is less than the sequence number of the evidence in the second evidence label. The higher the authority of the evidence, the earlier it is listed, and the smaller its sequence number.
[0073] Optionally, the response information may also include: third cited content and a third evidence tag corresponding to the third cited content. The third evidence tag includes an evidence sequence number tag and at least one authority tag. Wherein, when the priority of the authority tag included in the third evidence tag is higher than the authority tag included in the first evidence tag, the authority of the evidence corresponding to the third evidence tag is higher than the authority corresponding to the first evidence tag. The sequence number of the evidence sequence number tag in the third evidence tag is less than the sequence number of the evidence sequence number tag in the first evidence tag.
[0074] In this embodiment of the invention, authority and priority are in one-to-one correspondence. The higher the priority of the authority label, the higher the authority of the evidence.
[0075] For example, the authority label for the first piece of evidence is "expert consensus," indicating that the evidence corresponding to the first piece of evidence is based on expert consensus. The authority label for the third piece of evidence is "randomized controlled trial (RCT)," indicating that the evidence corresponding to the third piece of evidence is an RCT experiment. Based on the priority of the authority labels, the RCT label has a higher priority than the expert consensus label; therefore, the authority of the evidence corresponding to the third piece of evidence is higher than that of the first piece of evidence. This priority of the authority labels is highly consistent with the authority of medical evidence, aligning with the general understanding of medical professionals.
[0076] In some possible embodiments, in step 102, the computer device generates and displays response information based on the first information.
[0077] In some possible embodiments, in step 102, the computer device sends the first information to the server, the server generates response information based on the first information and sends the response information to the computer device, and the computer device displays the response information.
[0078] For example, the reply information is displayed in a waterfall layout.
[0079] Optionally, the authority label may include at least one of the following: StudyType label, source label, and journal ranking label.
[0080] For example, the StudyType tag includes the Meta Analysis tag, the Systematic Review tag, or the randomized controlled trial (RCT) tag.
[0081] For example, the source label may include a guide label or a site whitelist label.
[0082] For example, journal rating labels may include a first journal rating label, a second journal rating label, a third journal rating label, or a fourth journal rating label.
[0083] For example, the first journal rating label, the second journal rating label, the third journal rating label, or the fourth journal rating label can be represented by the Q1 label, the Q2 label, the Q3 label, and the Q4 label, respectively.
[0084] It should be noted that the above division of journals into four levels is only an example. In actual application, journal levels can be flexibly divided according to the actual situation, and different numbers of journal levels can be divided according to different classification strategies.
[0085] For example, Figure 3 This is a schematic diagram of the interface displaying the reply information of the first information in an embodiment of the present invention, such as... Figure 3 As shown, the first piece of information entered by the user was, "If a patient has been receiving hormone replacement therapy for a long time due to severe menopausal symptoms, is it safe to continue hormone replacement therapy in their 70s and 80s?" The response to this first piece of information contained three evidence tags. One evidence tag only contained an evidence number tag, number 3; the other two evidence tags contained both evidence number tags and authority tags. The authority tag corresponding to the evidence number tag for evidence number 1 was "Systematic Review," and the authority tag corresponding to the evidence number tag for evidence number 2 was "Q1." It can be seen that evidence containing an authority tag has higher authority than evidence without an authority tag.
[0086] Optionally, the authority label may also include a time label for the evidence. The time label indicates the time period of the evidence; for evidence of the same level, the more recent the evidence, the higher its authority.
[0087] For example, if the cited content corresponds to a medical guideline, then both an authority label and a time label will be displayed in the evidence label of that cited content. For instance, if the medical guideline is from 2025, then the time label will contain the text "2025".
[0088] This invention, by adding time tags to medical evidence-based responses, allows users to intuitively understand the time frame of the evidence, thus improving the user experience. For evidence of the same level, the more recent the evidence, the higher its authority. This enhances the authority and accuracy of medical evidence-based responses, further improving the user experience.
[0089] Optionally, the evidence corresponding to multiple cited contents can be sorted according to the priority of the pre-set authority labels to obtain the evidence number, and the evidence number corresponding to the cited contents can be displayed in the evidence number label.
[0090] For example, first select the highest priority authority label from all authority labels of each piece of evidence as the first authority label of the evidence. Then compare the priority of the first authority labels of different pieces of evidence. The higher the priority of the first authority label, the higher the authority of the evidence and the smaller the corresponding evidence number.
[0091] For example, the priority of the StudyType tag, source tag, and journal grade tag decreases in that order.
[0092] For example, the priority of medical guidelines, meta analysis tags, systematic review tags, and RTC tags decreases in that order.
[0093] For example, the priority of the guide label and the site whitelist label decreases in that order.
[0094] For example, the priority of the first journal rating label, the second journal rating label, the third journal rating label, and the fourth journal rating label decreases in that order.
[0095] For ease of understanding, priority can also be defined according to the following rules:
[0096] Evidence in the form of meta-analysis, systematic review, randomized controlled trials (RCTs), or medical guidelines has high priority; evidence in the form of non-randomized controlled trials (NRCTs), cohort studies, case-control studies, cross-sectional studies, literature reviews, or authoritative medical websites such as the Mayo Clinic has medium priority; evidence in the form of observational studies, case reports, animal trials, in vitro trials, expert consensus, scientific statements, or information from regular internet websites has low priority; or
[0097] Evidence from medical guidelines or drug instructions is of high priority; or
[0098] The evidence comes from a medical website and is of medium priority; or
[0099] Evidence from non-medical websites or other sources is given low priority.
[0100] Optionally, generating response information for the first information in step 102 includes: generating response information using a large language model based on the first information and a knowledge base, wherein the knowledge base contains multiple different pieces of evidence and authority labels corresponding to the evidence.
[0101] For example, the evidence could be a medical guideline, with the authority label containing the text "guideline". Another example is an RTC (Research and Development Center) document, with the authority label containing the text "RTC".
[0102] For example, a piece of evidence may contain one or more authority labels. If the evidence is both a medical guideline and a high-quality journal, it may contain two authority labels: one containing "guideline" and the other containing "high-quality journal".
[0103] For example, in the response information, all authority tags for a single cited content can be displayed, none can be displayed, or only a preset number of authority tags can be displayed. In the above embodiments of the present invention, the display of authority tags in the response information is selected based on authority level. However, in practice, this can be flexibly adjusted. Therefore, it can be determined whether to display authority tags for evidence corresponding to the cited content in the response information, and how many authority tags for evidence corresponding to the cited content should be displayed in the response information, based on the actual situation.
[0104] This invention constructs a knowledge base containing multiple different pieces of evidence and corresponding authority tags for each piece of evidence. This allows the response information to include authority tags for the evidence corresponding to the cited content, thereby improving the authority and accuracy of medical evidence-based response information and meeting the requirements of medical users.
[0105] Optionally, the evidence label may also include: evidence format label.
[0106] In this embodiment of the invention, the evidence format label is displayed in the response information, allowing users to intuitively understand the format of the evidence corresponding to the cited content, thereby improving the user experience.
[0107] Optionally, the format of the evidence format label and the evidence number label is a superscript. The superscript can be either a superscript or a subscript.
[0108] Optionally, the evidence format label is a PDF label. Figure 4 A flowchart of another medical evidence-based method provided in the embodiments of the present invention, such as Figure 4 As shown, after step 102, step 201 is also included.
[0109] Step 201: In response to the user's click on the evidence label with the PDF label, a PDF preview window pops up and displays the page of the evidence corresponding to the evidence label in the PDF preview window, and the cited content in the PDF preview window is highlighted.
[0110] In this embodiment of the invention, for evidence tags with PDF labels, clicking the evidence tag automatically pops up a PDF preview window displaying the evidence, making it convenient for users to view the evidence; at the same time, the cited content in the PDF preview window is highlighted, helping users to quickly locate the cited content in the PDF preview window and improving the user experience.
[0111] Optionally, such as Figure 4 As shown, after step 201, steps 202-203 are also included.
[0112] Step 202: In response to the user's selection of a portion of the evidence in the PDF preview window, display at least one relevant control, which may include a translation control, a copy control, or an AI search control.
[0113] Step 203: In response to the user's operation on the relevant controls, perform the operation corresponding to the relevant controls on a portion of the content.
[0114] In this embodiment of the invention, for evidence tags with PDF labels, clicking the evidence tag automatically pops up a PDF preview window displaying the evidence. The PDF preview window also provides relevant operations on the selected content, making it convenient for users to perform related operations on the evidence and improving the user experience.
[0115] Optionally, the response information may also include: an evidence list control. For example, such as... Figure 3 As shown, the evidence list control contains the text "All Evidence". Figure 5 A flowchart of another medical evidence-based method provided in the embodiments of the present invention, such as Figure 5 As shown, after step 102, step 301 is also included.
[0116] Step 301: In response to the user's click on the evidence list control, the evidence list pops up.
[0117] For example, Figure 6 This is a schematic diagram of the pop-up evidence list in an embodiment of the present invention, such as... Figure 6 As shown, after the user clicks the evidence list control, an evidence list pops up on the right side of the screen.
[0118] This invention improves user experience by adding an evidence list control to the reply information, providing users with an entry point to view the evidence list.
[0119] In some possible implementations, the evidence list is collapsed and automatically expanded when the user initiates a new evidence-based investigation and a complete evidence list is returned.
[0120] In some possible implementations, the evidence list is in an expanded state and is automatically updated when the user initiates new evidence-based activities and complete evidence list information is returned; or the evidence list is collapsed when the user clicks the close control of the evidence list; or the evidence list is collapsed before the first token of the response information is returned when the user enters new information without triggering evidence-based activities.
[0121] Optionally, the evidence list includes: evidence cards corresponding to the cited content, and the evidence cards contain an authority label for the evidence corresponding to the cited content.
[0122] In embodiments of the present invention, such as Figure 6 As shown, the evidence list displays relevant information about the evidence corresponding to the cited content in the form of evidence cards. By displaying the authority label of the evidence in the evidence cards, the authority label can reflect the authority of the cited content, thus improving the authority and accuracy of the medical evidence-based response information.
[0123] Optionally, such as Figure 6 As shown, the cited content comes from journals, and the evidence card also includes an IF label, which indicates the journal's impact value. This allows users to intuitively understand the journal's impact value, improving the user experience.
[0124] Optionally, the evidence card includes multiple authority labels for the evidence, which are ordered according to their priority.
[0125] In this embodiment of the invention, the evidence card can display all authority labels of the evidence corresponding to the cited content, such as... Figure 6 As shown, the evidence ranked first contains two authority labels: the Systematic Review label and the Q1 label. When an evidence card includes multiple authority labels, they are ranked according to their priority.
[0126] For example, multiple evidence labels are arranged in descending order of priority.
[0127] In this embodiment of the invention, the font, size, and format of each label in the evidence label, such as the authority label, time label, evidence serial number label, and evidence format label, are not specifically limited. They can be the same as the font, size, and format of the main text, or they can be presented in the form of superscript or subscript. The font, size, and format of each label can be the same or different.
[0128] Optionally, such as Figure 6 As shown, the evidence card also includes: the reason for including the evidence corresponding to the cited content. The reason for inclusion is generated by the large language model through analysis and summary of the first information and evidence.
[0129] In this embodiment of the invention, the evidence card also includes the reasons for including the evidence corresponding to the cited content. The reasons for inclusion are generated by the large language model through analysis and summary of the first information and evidence, which can improve the authority and accuracy of the medical evidence-based response information and enhance the user experience.
[0130] Optionally, multiple evidence cards in the evidence list can be sorted according to their authority.
[0131] For example, multiple evidence cards in the evidence list are sorted from strongest to weakest in terms of their authority.
[0132] In this embodiment of the invention, multiple evidence cards in the evidence list are sorted according to the authority of the evidence, so that users can see the evidence cards with stronger authority first, which can reflect the authority and accuracy of medical evidence-based results and improve the user experience.
[0133] Optionally, the evidence card may also include source information for the evidence, which includes the source (also known as a domain name) website name, icon, title, and supplementary information. Supplementary information may include the date, cited books, author information, and institutional information.
[0134] For example, the site name is displayed directly, and if it cannot be obtained, it can be displayed as "Reference Site".
[0135] For example, when the source of information is a medical guide, the site name is displayed as "Baichuan Special Library".
[0136] For example, the site's icon is displayed directly, and a fallback icon can be used when it cannot be obtained.
[0137] Optionally, such as Figure 6 As shown, the response information also includes: a first evidence control, which contains the serial number, name and source information of the first evidence, and the first evidence is the evidence ranked first in the evidence list.
[0138] In this embodiment of the invention, the computer device may also open the page of the first evidence in response to the user's operation on the first evidence control.
[0139] In this embodiment of the invention, the evidence ranked first in the evidence list has the highest authority. The response information includes a control for the evidence with the highest authority, allowing users to intuitively see the information of the most authoritative evidence and experience the authority and accuracy of the medical evidence-based response information, thereby improving the user experience.
[0140] Optionally, the serial number in the evidence serial number label is the serial number of the evidence card in the evidence list that corresponds to the cited content.
[0141] Optionally, such as Figure 6 As shown, the response information also includes: an expand control for the thought chain; the thought chain is displayed in a waterfall layout before the response information is displayed and collapses when the response information is started; the expand control is used to trigger the expansion of the thought chain in response to user actions.
[0142] Optionally, such as Figure 6 As shown, the reply message also includes: a sharing control. Figure 7 A flowchart of another medical evidence-based method provided in the embodiments of the present invention, such as Figure 7 As shown, after step 102, there are also steps 401-402.
[0143] Step 401: In response to the user's operation on the sharing control, select the first information and the reply result, and display the OK sharing control.
[0144] Step 402: In response to the user's operation on the share control, a share window pops up, which contains multiple share options.
[0145] Optionally, such as Figure 3 and Figure 6 As shown, the evidence tags in the response information are displayed in a highlighted manner. Highlighting evidence tags enhances the display effect, allowing users to quickly find the evidence tags and understand the relevant information of the cited content based on the evidence tags, thus improving the user experience.
[0146] Optionally, the response information may include a table, which is generated by the large language model through analysis and summarization based on the initial information and multiple pieces of evidence in the evidence list.
[0147] This invention provides a technical solution for a medical evidence-based method, which is applied to a human-computer dialogue application based on a large language model. The method includes: receiving first information related to medical evidence-based practice input by a user in the human-computer dialogue interface of the application; generating and displaying response information for the first information, wherein the response information includes: first cited content and a first evidence tag corresponding to the first cited content; wherein the first evidence tag corresponding to the first cited content includes an evidence sequence number tag and at least one authority tag, the authority tag indicating the authority of the evidence corresponding to the cited content. By introducing authority tags, this invention improves the authority and accuracy of the response information output by the large language model for medical evidence-based practice. Users can quickly identify the authority of the response information based on the authority tags, enhancing the user experience.
[0148] Figure 8 This is a schematic diagram of the structure of a medical evidence-based device provided in an embodiment of the present invention, as shown below. Figure 8 As shown, the device includes:
[0149] The receiving module 11 is used to receive first information related to medical evidence input by the user in the human-computer dialogue interface; the human-computer dialogue interface is the interface in the human-computer dialogue application based on the large language model.
[0150] Processing module 12 is used to generate response information for the first information;
[0151] Display module 13 is used to display the response information, the response information including:
[0152] The first cited content and the first evidence label corresponding to the first cited content;
[0153] The first evidence label corresponding to the first cited content includes an evidence number label and at least one authority label, wherein the authority label is used to indicate the authority of the evidence corresponding to the cited content.
[0154] In some possible embodiments, the response information further includes: second cited content and a second evidence tag corresponding to the second cited content, wherein the second evidence tag only includes the evidence sequence number tag, and the authority of the evidence corresponding to the first evidence tag including at least one authority tag is higher than the authority of the evidence corresponding to the second evidence tag which only includes the evidence sequence number tag.
[0155] In some possible embodiments, the response information further includes: second cited content and a second evidence tag corresponding to the second cited content, wherein the second evidence tag only includes the evidence sequence number tag, and the sequence number of the evidence sequence number tag in the first evidence tag is less than the sequence number of the evidence sequence number tag in the second evidence tag.
[0156] In some possible embodiments, the authority label includes at least one of a research type label, a source label, and a journal grade label.
[0157] In some possible embodiments, the authority label may also include a time label for the evidence.
[0158] In some possible embodiments, the processing module 12 is further configured to sort the evidence corresponding to the multiple cited contents according to the priority of the pre-set authority label to obtain the serial number of the evidence; the display module 13 is further configured to display the serial number of the evidence corresponding to the cited contents in the evidence serial number label.
[0159] In some possible embodiments, the processing module 12 is specifically used to generate the response information using the large language model based on the first information and the knowledge base; the knowledge base contains multiple different pieces of evidence and the authority labels corresponding to the evidence.
[0160] In some possible embodiments, the evidence label may further include an evidence format label.
[0161] In some possible embodiments, the format of the evidence format label and the evidence serial number label is a superscript.
[0162] In some possible embodiments, the evidence format label is a PDF label, and the processing module 12 is further configured to pop up a PDF preview window in response to the user's click operation on the evidence label having the PDF label;
[0163] The display module 13 is also used to display the page of the evidence corresponding to the evidence label in the PDF preview window, and the cited content in the PDF preview window is highlighted.
[0164] In some possible embodiments, the processing module 12 is further configured to respond to the user's selection of a portion of the evidence in the PDF preview window, and the display module 13 is further configured to display at least one related control, the related control including a translation control, a copy control, or a search control;
[0165] The processing module 12 is also configured to respond to the user's operation on the relevant control and perform an operation corresponding to the relevant control on the part of the content.
[0166] In some possible embodiments, the response information further includes: an evidence list control; the processing module 12 is further configured to pop up an evidence list in response to the user's click operation on the evidence list control; the evidence list includes: evidence cards corresponding to the cited content, and the evidence cards contain the authority label of the evidence corresponding to the cited content.
[0167] In some possible embodiments, the cited content comes from a journal, and the evidence card also includes an IF label indicating the impact value of the journal.
[0168] In some possible embodiments, the evidence card includes multiple authority labels for the evidence, and the multiple authority labels are ordered according to the priority of the authority labels.
[0169] In some possible embodiments, the evidence card further includes: a reason for including the evidence corresponding to the cited content, the reason for inclusion being generated by the large language model through analysis and summarization of the first information and the evidence.
[0170] In some possible embodiments, the evidence cards in the evidence list are ordered according to the authority of the evidence.
[0171] In some possible embodiments, the response information further includes: a first evidence control, the first evidence control containing the serial number, name and source information of a first piece of evidence, the first piece of evidence being the first piece of evidence in the evidence list;
[0172] The processing module 12 is also used to respond to the user's operation on the first evidence control, and the display module 13 is used to open the page of the first evidence.
[0173] In some possible embodiments, the response information further includes: a mind chain expansion control; the mind chain is displayed in a waterfall layout before the response information is displayed and collapses when the response information is started to be displayed; the expansion control is used to trigger the expansion of the mind chain in response to a user's operation.
[0174] In some possible embodiments, the evidence label is displayed in a highlighted manner.
[0175] In the technical solution provided by this invention, first information related to medical evidence-based practice is received from a user in a human-computer dialogue interface. The human-computer dialogue interface is an interface within a human-computer dialogue application based on a large language model. A response to the first information is generated and displayed. This response includes first cited content and a first evidence tag corresponding to the first cited content. The first evidence tag corresponding to the first cited content includes an evidence sequence number tag and at least one authority tag, whereby the authority tag indicates the authority of the evidence corresponding to the cited content. By introducing authority tags, this invention improves the authority and accuracy of the response information output by the large language model for medical evidence-based practice. Users can quickly identify the authority of the response information based on the authority tags, enhancing the user experience.
[0176] This application provides a storage medium that includes a stored program, wherein the program controls the device where the storage medium is located to execute the above-described method when it is running.
[0177] This application provides a computer device including a memory and a processor. The memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions. When the program instructions are loaded and executed by the processor, they implement the steps of the above-described method.
[0178] Figure 9 This application provides a schematic diagram of the structure of a computer device, as shown in the embodiment of the present application. Figure 9 As shown, the computer device 20 includes a processor 21, a memory 22, and a computer program 23 stored in the memory 22 and executable on the processor 21. When the computer program 23 is executed by the processor 21, it implements the medical evidence-based methods in the embodiments. To avoid repetition, these will not be described in detail here.
[0179] Computer device 20 includes, but is not limited to, processor 21 and memory 22. Those skilled in the art will understand that... Figure 9 This is merely an example of computer device 20 and does not constitute a limitation on computer device 20. It may include more or fewer components than shown, or combine certain components, or different components. For example, computer device 20 may also include input / output devices, network access devices, buses, etc.
[0180] The processor 21 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.
[0181] The memory 22 can be an internal storage unit of the computer device 20, such as a hard disk or RAM of the computer device 20. The memory 22 can also be an external storage device of the computer device 20, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the computer device 20. Furthermore, the memory 22 can include both internal and external storage units of the computer device 20. The memory 22 is used to store computer programs and other programs and data required by the computer device 20. The memory 22 can also be used to temporarily store data that has been output or will be output.
[0182] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0183] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.
[0184] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0185] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.
[0186] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0187] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.
Claims
1. A medical evidence-based approach, applied to a human-computer dialogue application based on a large language model, characterized in that, The method includes: Receive first information related to medical evidence-based practice input by the user in the human-computer dialogue interface of the human-computer dialogue application; Generate and display a response message for the first information, the response message including: The first cited content and the first evidence label corresponding to the first cited content; The first evidence label corresponding to the first cited content includes an evidence number label and at least one authority label, wherein the authority label is used to indicate the authority of the evidence corresponding to the cited content.
2. The method according to claim 1, characterized in that, The response information also includes: second cited content and a second evidence tag corresponding to the second cited content. The second evidence tag only includes the evidence sequence number tag. The authority of the evidence corresponding to the first evidence tag that includes at least one authority tag is higher than the authority of the evidence corresponding to the second evidence tag that only includes the evidence sequence number tag.
3. The method according to claim 1, characterized in that, The response information also includes: second cited content and a second evidence tag corresponding to the second cited content. The second evidence tag only includes the evidence sequence number tag, wherein the sequence number of the evidence sequence number tag in the first evidence tag is less than the sequence number of the evidence sequence number tag in the second evidence tag.
4. The method according to claim 1, characterized in that, The authority label includes at least one of the following: research type label, source label, and journal level label.
5. The method according to claim 1, characterized in that, The authority label also includes a time label for the evidence.
6. The method according to any one of claims 1-4, characterized in that, The method further includes: The evidence is ordered according to the priority of the pre-set authority label to obtain the serial number of the evidence corresponding to the multiple cited contents; The evidence number label displays the number of the evidence corresponding to the cited content.
7. The method according to claim 1, characterized in that, The response information that generates the first information includes: Based on the first information and the knowledge base, the response information is generated using the large language model; the knowledge base contains multiple different pieces of evidence and the authority labels corresponding to the evidence.
8. The method according to any one of claims 1-5 and 7, characterized in that, The evidence label also includes: evidence format label.
9. The method according to claim 8, characterized in that, The format of the evidence format label and the format of the evidence serial number label are superscripts.
10. The method according to claim 8, characterized in that, The evidence format label is a PDF label, and the method further includes: In response to the user's click operation on the evidence tag with the PDF tag, a PDF preview window pops up and displays the page of the evidence corresponding to the evidence tag in the PDF preview window, and the cited content in the PDF preview window is highlighted.
11. The method according to claim 10, characterized in that, The method further includes: In response to the user's selection of a portion of the evidence in the PDF preview window, at least one relevant control is displayed, including a translation control, a copy control, or a search control; In response to the user's operation on the relevant control, the operation corresponding to the relevant control is performed on the part of the content.
12. The method according to claim 1, characterized in that, The response information also includes: an evidence list control; the method also includes: In response to the user's click on the evidence list control, an evidence list pops up; the evidence list includes: evidence cards corresponding to the cited content, and the evidence cards contain the authority label of the evidence corresponding to the cited content.
13. The method according to claim 12, characterized in that, The cited content comes from a journal, and the evidence card also includes an IF label, which is used to indicate the impact value of the journal.
14. The method according to claim 12, characterized in that, The evidence card includes multiple authority labels for the evidence, and the multiple authority labels are ordered according to their priority.
15. The method according to claim 12, characterized in that, The evidence card also includes: a reason for including the evidence corresponding to the cited content, wherein the reason for inclusion is generated by the large language model through analysis and summary of the first information and the evidence.
16. The method according to claim 12, characterized in that, The evidence cards in the evidence list are sorted according to their authority.
17. The method according to claim 16, characterized in that, The response information also includes: a first evidence control, which contains the serial number, name and source information of the first evidence, wherein the first evidence is the evidence ranked first in the evidence list; The method further includes: In response to the user's operation on the first evidence control, the page of the first evidence is opened.
18. The method according to claim 1, characterized in that, The response information also includes: a mind chain expansion control; the mind chain is displayed in a waterfall layout before the response information is displayed and collapses when the response information is started to be displayed; the expansion control is used to trigger the expansion of the mind chain in response to user operation.
19. The method according to claim 1 or 12, characterized in that, The evidence tags are displayed in a highlighted manner.
20. A medical evidence-based device, characterized in that, include: The receiving module is used to receive the first information related to medical evidence-based practice input by the user in the human-computer dialogue interface; The human-computer dialogue interface is the interface in a human-computer dialogue application based on a large language model. The processing module is used to generate the response information for the first information; A display module is used to display the response information, which includes: The first cited content and the first evidence label corresponding to the first cited content; The first evidence label corresponding to the first cited content includes an evidence number label and at least one authority label, wherein the authority label is used to indicate the authority of the evidence corresponding to the cited content.
21. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the storage medium is located to perform the method according to any one of claims 1 to 19.
22. A computer device comprising a memory and a processor, the memory for storing information including program instructions, and the processor for controlling the execution of the program instructions, characterized in that, When the program instructions are loaded and executed by the processor, they implement the steps of the method described in any one of claims 1 to 19.
