Clinical research information acquisition method and related device
Through the pre-trained clinical research model, the target question text is determined based on the goal needs of clinical research, and information is obtained based on the answer types, which solves the problems of weak knowledge and difficult data processing in clinical research, and improves research efficiency and accuracy.
Patent Information
- Application Number
- CN202411907723.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-23
- Publication Date
- 2025-04-18
AI Technical Summary
The lack of systematic learning by scientific researchers in clinical research leads to weak knowledge mastery, difficult ideas and difficult to process large-scale complex data, which in turn leads to difficulties in research development and difficult production of results.
Provide a clinical research information acquisition method. Through a pre-trained clinical research model, the target question text is determined based on the target needs, and the answer type is judged to determine the information acquisition method, so as to achieve problem guidance and information generation for the research link.
It reduces the complexity and difficulty of clinical research, improves research efficiency, determines the information acquisition method through question text guidance ideas and combines answer types, realizes multi-data source analysis, and enhances the accuracy and reliability of the research.
Smart Images
Figure CN119851957A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of clinical research technology, and in particular to a method for acquiring clinical research information and related devices. Background Art
[0002] Clinical research is a key branch of the medical field. It can connect basic research with clinical practice and is crucial to improving treatment, enhancing service quality and promoting medical progress. It requires researchers to have both medical knowledge and data analysis capabilities to ensure the rationality, feasibility and reliability of the research, accelerate transformation and innovation, and improve the level of medical development.
[0003] At present, clinical research is generally handled manually, but scientific researchers may lack systematic training in clinical research, resulting in weak knowledge of research, difficulty in conceiving scientific research ideas, and difficulty in coping with large-scale and complex data processing needs. This in turn makes it difficult to carry out clinical research and produce research results. Summary of the invention
[0004] Based on the above problems, the present application provides a method and related devices for obtaining clinical research information, aiming to reduce the complexity of clinical research and improve the efficiency of clinical research.
[0005] The embodiments of the present application disclose the following technical solutions:
[0006] In a first aspect, an embodiment of the present application provides a method for obtaining clinical research information, the method comprising:
[0007] Obtain the target needs of the target subjects; the target needs indicate the target research links of the clinical research;
[0008] Through the pre-trained clinical research model, the target question text is determined according to the target needs; the target question text indicates the questions related to the target research link; the clinical research model is trained based on the research database and the research corpus, the research database includes multiple annotated medical data, and the research corpus includes question and answer corpora corresponding to multiple research links;
[0009] Determine whether the answer type corresponding to the target question text is a preset type;
[0010] If the answer type is a preset type, the first clinical research information is determined according to the target question text through the clinical research model; the type of the first clinical research information is a preset type;
[0011] If the answer type is not a preset type, the target interface is determined according to the answer type, and through the target interface, the second clinical research information is determined according to the target question text, and the type of the second clinical research information is the answer type.
[0012] Optionally, when the target research link is the second research link, the first clinical research information is determined according to the target question text through the clinical research model, including:
[0013] Through the clinical research model, the first clinical research information is determined according to the historical clinical research information and the question text; the historical clinical research information is the clinical research information corresponding to the first research link; the first research link is associated with the second research link;
[0014] Correspondingly, determining the second clinical research information according to the target question text through the target interface includes: determining the second clinical research information according to the target question text and historical clinical research information through the target interface.
[0015] Optionally, before determining the first clinical research information according to the historical clinical research information and the question text through the clinical research model, the method further includes:
[0016] The clinical research information corresponding to the first research phase is cut and processed to obtain historical clinical research information.
[0017] Optionally, the first clinical research information is determined according to the target question text through the clinical research model, including:
[0018] Obtain input data corresponding to the target research link;
[0019] Through the clinical research model, the first clinical research information is determined according to the target question text and the input data.
[0020] Optionally, a scientific research corpus is obtained by:
[0021] Obtain a clinical research question set; the clinical research question set includes multiple clinical research questions;
[0022] Classifying and labeling each clinical research problem in the clinical research problem set to obtain multiple subsets; the types and labeling results of each clinical research problem in the subsets are the same;
[0023] Determine a question template corresponding to each subset; the question template is obtained based on the clinical research questions in the corresponding subset;
[0024] Establish a scientific research corpus based on multiple question-answer templates.
[0025] Optionally, the second clinical research information is determined according to the target question text through the target interface, including:
[0026] Calling a target tool or a target database through a target interface;
[0027] The target question text is processed according to the target tool or the target database to obtain the second clinical research information.
[0028] Optionally, it also includes: after determining the clinical research information, using the clinical research information to train the clinical scientific research model; the clinical research information is the first clinical research information or the second clinical research information.
[0029] In a second aspect, an embodiment of the present application provides a clinical research information acquisition device, the device comprising:
[0030] An acquisition module is used to acquire target requirements of target subjects; target requirements indicate target research links of clinical research;
[0031] The question text determination module is used to determine the target question text according to the target requirements through the pre-trained clinical research model; the target question text indicates the questions related to the target research link; the clinical research model is trained based on the research database and the research corpus, the research database includes multiple annotated medical data, and the research corpus includes question and answer corpora corresponding to multiple research links;
[0032] A judgment module is used to judge whether the answer type corresponding to the target question text is a preset type;
[0033] A first information acquisition module, for determining first clinical research information according to the target question text through a clinical research model if the answer type is a preset type; the type of the first clinical research information is a preset type;
[0034] The second information acquisition module is used to determine the target interface according to the answer type if the answer type is not a preset type, and determine the second clinical research information according to the target question text through the target interface, and the type of the second clinical research information is not a preset type.
[0035] In a third aspect, an embodiment of the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the clinical research information acquisition method as described in any one of the first aspects is implemented.
[0036] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium. When the instructions are executed on a terminal device, the terminal device executes the clinical research information acquisition method as described in any one of the first aspects.
[0037] Compared with the prior art, this application has the following beneficial effects:
[0038] The clinical research information acquisition method provided in the embodiment of the present application determines the target question text according to the target requirements of the target research link of the clinical research through a pre-trained clinical scientific research model, and further determines whether the answer type corresponding to the target question text is a preset type. If the answer type is the preset type, the first clinical research information is determined according to the target question text through the clinical scientific research model; if the answer type is not the preset type, the target interface is determined according to the answer type, and the second clinical research information is determined according to the target question text through the target interface.
[0039] Among them, the target question text indicates the questions related to the target research link; the clinical research model is trained based on the scientific research database and the scientific research corpus, the scientific research database includes multiple annotated medical data, the scientific research corpus includes question and answer corpora corresponding to multiple research links, the type of the first clinical research information is a preset type, and the type of the second clinical research information is an answer type.
[0040] In this way, this application uses a pre-trained clinical research model to determine the question text for the research link according to the target requirements, and then generates corresponding clinical research information based on the question text. On the one hand, the question text can provide users with guiding ideas and reduce the difficulty of clinical research. On the other hand, it can determine whether it is necessary to call tools or databases outside the model for processing based on the answer type corresponding to the question text, which can realize multi-data source analysis, reduce the difficulty of clinical research, and improve the efficiency of clinical research. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0042] Figure 1 A flowchart of a method for obtaining clinical research information provided in an embodiment of the present application;
[0043] Figure 2 A schematic diagram of a clinical research process provided in an embodiment of the present application;
[0044] Figure 3 A schematic diagram of a process for establishing a scientific research corpus provided in an embodiment of the present application;
[0045] Figure 4 A modular schematic diagram of a target question text provided in an embodiment of the present application;
[0046] Figure 5A schematic diagram of a clinical research information acquisition process provided in an embodiment of the present application;
[0047] Figure 6 A schematic diagram of the structure of a clinical research information acquisition device provided in an embodiment of the present application;
[0048] Figure 7 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0049] Since the process of clinical research requires researchers to have solid medical knowledge and data management and analysis capabilities to ensure the rationality, feasibility and reliability of clinical research, researchers can use multiple clinical research-related tools to complete clinical research work on the basis of manual clinical research. For example, tools related to clinical research can be medical big data platforms, large language models, etc.
[0050] The medical big data platform is a platform that integrates data storage, data extraction, data preprocessing, and statistical analysis functions, which enables scientific researchers to extract, count, and analyze medical data more quickly.
[0051] The Large Language Model (LLM) is a deep learning model trained on a large amount of text data. The Large Language Model can not only generate natural language text, but also deeply understand the meaning of the text and handle various natural language tasks such as text summarization, question answering, translation, etc.
[0052] However, the tools related to clinical research provided by relevant technologies still have the following problems:
[0053] (1) In the early stages of clinical research, researchers need a lot of time to manually explore research directions, but the medical big data platform cannot query, summarize, and mine scientific research literature in medical databases outside the platform, which increases the difficulty of clinical research;
[0054] (2) The data preprocessing function of the medical big data platform is limited and cannot fully meet the requirements of clinical research. For example, for the processing of complex variables, the exported data still needs to be manually cleaned and processed, which reduces the efficiency of clinical research;
[0055] (3) The tools in the relevant technologies generally lack text translation and polishing functions. In the later stages of clinical research, researchers need to spend a lot of time translating and proofreading the compiled articles to ensure the rigor and accuracy of the wording, which increases the difficulty of clinical research.
[0056] (4) The tools of related technologies can only be used to conceive based on the existing research needs of researchers, and cannot guide and help researchers to expand their ideas and give research suggestions. For researchers with less research experience, there may be incomplete or non-standard ideas. If they cannot be discovered and corrected in time, it may cause errors in the results of clinical research, requiring more time and energy to adjust the research plan, reducing the efficiency of clinical research;
[0057] (5) Large language models are usually general models, and their application in the medical vertical field is still immature. Especially in clinical research, it takes a lot of time to customize and optimize large language models to meet the needs of researchers, which also increases the difficulty of clinical research.
[0058] It should be understood that although the tools related to clinical research in the relevant technologies can assist researchers in conducting clinical research, they will still increase the difficulty of clinical research and reduce the efficiency of clinical research.
[0059] In order to solve the above problems, the embodiment of the present application provides a method for obtaining clinical research information and a related device. The method includes: determining a target question text according to the target requirements of the target research link of the clinical research through a pre-trained clinical research model, further determining whether the answer type corresponding to the target question text is a preset type, and if the answer type is a preset type, determining the first clinical research information according to the target question text through the clinical research model; if the answer type is not a preset type, determining the target interface according to the answer type, and determining the second clinical research information according to the target question text through the target interface.
[0060] Among them, the target question text indicates the questions related to the target research link; the clinical research model is trained based on the scientific research database and the scientific research corpus, the scientific research database includes multiple annotated medical data, the scientific research corpus includes question and answer corpora corresponding to multiple research links, the type of the first clinical research information is a preset type, and the type of the second clinical research information is an answer type.
[0061] In this way, this application uses a pre-trained clinical research model to determine the question text for the research link according to the target requirements, and then generates corresponding clinical research information based on the question text. On the one hand, the question text can provide users with guiding ideas and reduce the difficulty of clinical research. On the other hand, it can determine whether it is necessary to call tools or databases outside the model for processing based on the answer type corresponding to the question text, which can realize multi-data source analysis, reduce the difficulty of clinical research, and improve the efficiency of clinical research.
[0062] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0063] See also Figure 1 , which is a flow chart of a method for obtaining clinical research information provided in an embodiment of the present application, combined with Figure 1 As shown, the clinical research information acquisition method provided in the embodiment of the present application may include:
[0064] S101: Obtain target requirements of a target object.
[0065] Target subjects refer to the subjects on whom clinical research is needed, such as researchers, doctors, etc.
[0066] Target demand refers to the clinical research demand formulated by the target object according to the actual situation. For example, if the hot spot type of cold-like diseases needs to be determined, the target demand can indicate the target research link of the clinical research. Figure 2 As shown, a clinical research process may include but is not limited to research hotspot mining, research program design, data processing and statistics, and paper polishing and adjustment.
[0067] Research hotspots refer to topics or directions in the medical field that are currently receiving widespread attention, are actively researched, and are of great significance. These research hotspots usually reflect the latest progress, future trends, and potential innovations in the medical field. The research hotspot mining link is used to mine research points or research directions from various medical data after sorting and summarizing them.
[0068] Research plan refers to the clinical research plan formulated based on research hotspots. The research plan design phase is used to design and organize the research plan based on research hotspots.
[0069] Data processing and statistics refers to the process of processing and counting the data involved in a determined research plan. The data processing and statistics link is used to process and count the corresponding data according to the research plan to obtain the corresponding results of the research plan.
[0070] Paper polishing and adjustment means to enrich the semantics and adjust the wording of the paper written based on the results of the research plan. The paper polishing and adjustment link is used to polish and adjust the paper generated based on the results of the plan.
[0071] In one possible implementation, the target requirements can be obtained by:
[0072] Receive the target object's login information and historical usage records, and analyze user needs; recommend research directions that may be of interest to the target object (such as current research progress on the disease of concern, high-scoring research literature, etc.) to enhance the user's scientific research interest; further, analyze the user's needs based on the research direction text or clinical research demand text selected by the target object, and determine the target research link corresponding to the user's needs; and then determine the target needs based on the target research link.
[0073] S102: Determine the target question text according to target requirements through the pre-trained clinical research model.
[0074] The clinical research model refers to a model used to generate clinical research information corresponding to the target demand based on the target demand. The clinical research model can be trained based on the research database and the research corpus.
[0075] In a possible implementation, the pre-trained clinical research model is a large language model containing professional knowledge in the field of clinical research, which is obtained by training with data from a scientific research database and a scientific research corpus, based on a general large language model (such as ChatGLM, Qwen, LLaMA, baichuan and other open source large language models). It should be noted that, since the general large language is highly versatile but less specialized in specific professional fields, it is necessary to use data related to clinical research to construct a scientific research corpus in the embodiments of the present application, and to pre-train the basic general large language model, thereby obtaining a pre-trained clinical research model.
[0076] The scientific research database includes multiple annotated medical data, such as annotated patient electronic medical records, laboratory tests, de-privacy data of imaging examination reports, clinical data, disease diagnosis and treatment guidelines, medical terminology, clinical literature, etc. The scientific research database uses rich medical data to help the clinical research language model have professional medical knowledge and be able to answer questions related to clinical research and give corresponding answers.
[0077] The scientific research corpus includes question and answer corpora corresponding to multiple research links. The scientific research corpus provides question templates that are more in line with the clinical research process, making it more in line with the research needs and preferences of the target objects, thereby improving the interactive experience and improving the accuracy of the clinical research model. At the same time, a prompt strategy is designed for the clinical research model in the form of a question and answer template to provide targeted prompts (i.e. prompts related to the research link) for the target objects and the clinical research model, so that the clinical research model has better question and answer performance in the field of clinical research.
[0078] It should be understood that the scientific research corpus can be classified according to the four research links of clinical research, sort out the research problem corpus of each research link, and annotate the research problem corpus with keywords to construct a scientific research corpus. The corpus in the scientific research corpus can be used as prompt information to provide precise expressions to make the answers of the clinical research model more accurate. The scientific research corpus also has the ability to associate context, so that the clinical research model trained and fine-tuned by the corpus in the scientific research corpus has the ability of context, connecting each round of dialogue, guiding users to sort out research problems, and at the same time guiding the next research link, promoting the clinical research process, and reducing the difficulty of scientific research.
[0079] In some possible implementations, question templates can also be introduced during the training of clinical research models. Question templates refer to question templates for corresponding research links. Each question template is obtained based on the question and answer corpus of the corresponding research link, and the target question and answer template is the question and answer template corresponding to the target research link. For example, question templates for literature retrieval and summary, question templates for various statistical methods for statistical analysis of research data, question templates for results, etc.
[0080] It should be understood that by using the question template to train the clinical research model, the format of the research results produced by the clinical research model in the clinical research process can be standardized. For example, if the research link is research hotspot mining, the pre-trained clinical research model can summarize the retrieved research literature, and the format of the summary returned by the pre-trained clinical research model is: research purpose: XXX, research object: XXX, research method: XXX, research conclusion: XXX.
[0081] In the fields of linguistics, natural language processing (NLP), information retrieval, question-answering systems, etc., question corpus (or "question sentences", "clinical research questions", etc.) usually refers to a series of question sentences or a collection of sentences used in research. Question corpus can be manually written or extracted from existing texts (such as web pages, books, conversation records, etc.). Among them, the role of question-answering corpus is similar to that of scientific research corpus. On this basis, question-answering corpus can also enable clinical research models to learn answering patterns and correctly understand and answer corresponding questions by annotating questions and answers.
[0082] In one possible implementation, combining Figure 3 As shown, the scientific research corpus in the embodiment of the present application can be obtained by the following methods:
[0083] S201: Obtain a set of clinical research questions.
[0084] Among them, the clinical research problem set includes multiple clinical research problems. Clinical research problems refer to problems arising in clinical research, such as "clinical research on the therapeutic effect and safety of a new drug on a specific disease", "comparative analysis of the efficacy of different surgical methods in the treatment of a certain tumor disease", etc.
[0085] S202: Classify and label each clinical research problem in the clinical research problem set to obtain multiple subsets.
[0086] Among them, classification refers to dividing clinical research problems of the same type into the same subset, and labeling refers to labeling the clinical research problems with corresponding types or other identifications that the types and labeling results of each clinical research problem in the subset are the same.
[0087] For example, the set of clinical research questions includes: "Evaluation of the therapeutic effects of different treatment plans for a specific disease", "Comparative analysis of the efficacy of different surgical methods in the treatment of a certain tumor disease", "Exploration of related risk factors for the onset of a specific disease", and "Cross-sectional study of the epidemiological characteristics of a specific disease".
[0088] Among them, "Evaluation of the therapeutic effect of different treatment plans for a specific disease" and "Comparative analysis of the efficacy of different surgical methods in the treatment of a certain tumor disease" belong to the prognosis analysis type of problems; "Exploring the relevant risk factors for the onset of a specific disease" belongs to the etiology analysis type of problems; "Cross-sectional study of the epidemiological characteristics of a specific disease" belongs to the disease distribution type of problems. Therefore, these three types of problems can be divided into three subsets, and each clinical research problem can be labeled.
[0089] S203: Determine the question template corresponding to each subset.
[0090] The question template is obtained based on the clinical research questions in the corresponding subset. For example, each clinical research question is cleaned to obtain multiple processed clinical research questions, and the multiple processed clinical research questions are used as question templates. The method of determining the question template is not limited here.
[0091] S204: Establish a scientific research corpus based on multiple question and answer templates.
[0092] It should be understood that question templates provide a structured framework that makes the creation and presentation of questions more efficient and consistent. Question templates ensure that the format, structure, and language style of the question texts that are subsequently determined remain consistent, thereby improving the speed and accuracy of clinical research processing. In addition, question templates usually contain some key information points and structural elements that help researchers better understand the intent and context of the question. Through a templated approach, it is easier to extract the key information of the question and effectively process and analyze it.
[0093] The target question text indicates the question related to the target research segment.
[0094] It should be understood that in one possible implementation, the question template can provide sufficient contextual information for the clinical research model to help the clinical research model understand the background of the target question text. At the same time, it can also provide reminders for key language clues to guide researchers to determine the target question text corresponding to the target needs based on the question template.
[0095] As an example, assuming the question template is: "What is the purpose of the study?", "Who are the research subjects?", "How will you collect data?", "What variables do you plan to analyze?", "What are the expected research results?", then the corresponding target question text can be determined based on the question template: The purpose of the study is to evaluate the effect of a new diabetes drug on blood sugar control. The research subjects are 50 patients aged 40-60 with type 2 diabetes. Data needs to be collected through blood sugar monitoring and questionnaires. It is planned to analyze blood sugar levels, medication frequency and patients' lifestyle variables. The expected research result is that the new drug can significantly reduce patients' blood sugar levels.
[0096] That is, by conducting guided question-and-answer sessions using question-and-answer templates, researchers can be prompted to dig deep into the problem in the form of a thought chain, and step by step, researchers can clarify their clinical research needs.
[0097] In one possible implementation, combining Figure 4 As shown, the process of determining the target question text can be implemented through a demand receiving module 401 , a model parsing module 402 , a result display module 403 and an intelligent recommendation module 404 .
[0098] The requirement receiving module 401 can collect the requirement text of the target object (ie, the target requirement) and transmit the requirement text to the model parsing module 402 .
[0099] The model analysis module 402 uses the clinical research model, which determines the corresponding question template according to the target requirements, and then sends the question template to the result display module 403.
[0100] The result display module 403 displays the question template through the front-end display page, and prompts researchers with the content and precautions of the next stage of research work through the intelligent recommendation module 404.
[0101] In this way, researchers can generate target question text based on the question template displayed by the result display module 403 and the content prompted by the intelligent recommendation module 404.
[0102] S103: Determine whether the answer type corresponding to the target question text is a preset type.
[0103] Answer type refers to the type of answer determined by the target question text, such as text, Excel table, chart, picture, code, etc. Preset type refers to the type supported by the function of the clinical research model, such as text, Excel table, etc.
[0104] It should be understood that in order to ensure the performance of the clinical research model, generally not many functions are set. Therefore, when the answer type corresponding to the target question text is not a preset type, it is necessary to generate clinical research information in step S106.
[0105] S104: If the answer type is a preset type, the first clinical research information is determined according to the target question text through the clinical research model.
[0106] Among them, the type of the first clinical research information is a preset type.
[0107] It should be understood that if the answer type corresponding to the target question text is a preset type supported by the function of the clinical research model, then the clinical research model can be directly used to obtain the first clinical research information corresponding to the target question text. As an example, assuming that the preset type is text and the answer type corresponding to the target question text A is text, then the target question text A can be directly input into the clinical research model to obtain the output type of the first clinical research information B as text.
[0108] S105: If the answer type is not a preset type, determine the target interface according to the answer type, and determine the second clinical research information according to the target question text through the target interface.
[0109] The type of the second clinical research information is not a preset type, but an answer type corresponding to the target question text.
[0110] It should be understood that if the answer type corresponding to the target question text is not a preset type supported by the function of the clinical research model, the target interface corresponding to the answer type and the tool or database that can implement the answer type can be processed to obtain the second clinical research information. As an example, assuming that the preset type is text and the answer type corresponding to the target question text C is XML text, the target interface W corresponding to the XML text can be determined, and the corresponding function can be called through the target interface W to obtain the second clinical research information D of the type of XML text corresponding to the target question text C.
[0111] In a possible implementation, step S105 may include:
[0112] A1: Call the target tool or target database through the target interface.
[0113] It should be understood that if the answer type is not a preset type, the clinical research model can call the target tool or target database through the target interface (such as API interface) to achieve query or analysis functions. The target tool can be a medical big data platform, a large language model, etc., and the target database can be a medical database, etc., which are not specifically limited here.
[0114] A2: The target question text is processed according to the target tool or target database to obtain the second clinical research information.
[0115] It should be understood that in the embodiments of the present application, by processing through target tools or target databases other than the model, multi-data source analysis can be achieved, the difficulty of clinical research can be reduced, and the efficiency of clinical research can be improved.
[0116] Among them, if the target tool is a medical big data platform, the results (second clinical research information) output by the called function will be displayed on the corresponding interface of the medical big data platform, and the results and parameters of calling the medical big data platform will be summarized.
[0117] It should be understood that clinical research information means the results corresponding to the research links indicated by the target requirements. For example, if the research link is the research hotspot mining link, then the clinical research information is the research hotspot.
[0118] In an embodiment of the present application, a pre-trained clinical research model is used to determine the question text for the research link according to the target requirements, and then the corresponding clinical research information is generated based on the question text. On the one hand, the question text can provide guidance for users and reduce the difficulty of clinical research. On the other hand, the answer type corresponding to the question text can be combined to determine whether it is necessary to call tools or databases outside the model for processing, which can realize multi-data source analysis, reduce the difficulty of clinical research, and improve the efficiency of clinical research.
[0119] Based on the clinical research information acquisition method provided in the above embodiment, in a possible implementation, after the clinical research information is determined, the clinical research model can also be trained using the clinical research information; the clinical research information is the first clinical research information or the second clinical research information.
[0120] It should be understood that by annotating the target question text and its clinical research information and establishing a positive feedback mechanism, and using the clinical research information for training, on the one hand, the clinical research information and the corresponding target question text can be stored in the scientific research corpus to enhance the data quality and richness of the scientific research corpus. On the other hand, the question template can be improved, thereby improving the accuracy and reliability of the output results of the clinical research model.
[0121] Based on the clinical research information acquisition method provided in the above embodiment, in a possible implementation, when the target research link is the second research link, step S104 may include: determining the first clinical research information according to the historical clinical research information and the question text through the clinical scientific research model.
[0122] Among them, the historical clinical research information is the clinical research information corresponding to the first research link; the first research link is related to the second research link.
[0123] Correspondingly, step S105 may include: determining the second clinical research information according to the target question text and the historical clinical research information through the target interface.
[0124] It should be understood that the clinical research process can include research hotspot mining, research program design, data processing and statistics, and paper polishing and adjustment. The various links can be connected in series. For example, if the target requirement is to output a research program, then it is necessary to first determine the clinical research information corresponding to the research hotspot mining link, and use the clinical research information corresponding to the research hotspot mining link as the research program design link. Therefore, when determining the clinical research information corresponding to the second research link, it is necessary to combine the historical clinical research information of the first research link to reduce the difficulty of the model in the training stage, improve the reusability of the model, reduce the complexity of clinical research, and improve the efficiency of clinical research.
[0125] In a possible implementation, the historical clinical research information can be obtained based on the clipping of the clinical research information corresponding to the first research link. It should be understood that since the clinical research information corresponding to the first research link may not contain all the information required by the user, the clinical research information corresponding to the first research link can be clipped to obtain the clinical research information required by the user as the historical clinical research information, thereby improving the accuracy and efficiency of the clinical research.
[0126] It should be noted that the target research link can be the third research link or the fourth research link. The clinical research model can execute the above processes in sequence according to the order of the research links, and use the output clinical research information as the input of the next research link. The relevant processes of the third research link and the fourth research link can refer to the processing process of the target research link being the second research link, which will not be repeated here.
[0127] Based on the clinical research information acquisition method provided in the above embodiment, in another possible implementation, when the target research link is the second research link, step S104 may include: obtaining input data corresponding to the target research link, and determining the first clinical research information according to the target problem text and the input data through the clinical scientific research model.
[0128] Correspondingly, step S105 may include: determining the second clinical research information according to the target question text and the input data through the target interface.
[0129] It should be understood that in the embodiment of the present application, the pre-trained clinical research model can also receive data input by the user (i.e., input data), and then determine the corresponding clinical research information based on the input data and the target question text. For example, the target research link is the third research link, and the input data input by the user is "design for xx disease", and then the clinical research model can generate the data required for "design for xx disease" based on "design for xx disease" and the target question text as clinical research information.
[0130] In this way, different personalized options can be provided to users, the accuracy of clinical research information can be enhanced, user needs can be met, and user experience can be improved.
[0131] Based on the clinical research information acquisition method provided in the above embodiment, see Figure 5 , which is a schematic diagram of a clinical research information acquisition process provided in an embodiment of the present application, combined with Figure 5 As shown, another clinical research information acquisition process provided by the embodiment of the present application may include:
[0132] Step 1: Enter the target needs of the target audience.
[0133] The target requirement may be in text form, which may be obtained by converting speech to text, extracting text from images, converting PDF to text, etc., which is not specifically limited here.
[0134] Step 2: Use the pre-trained clinical research model to determine the target question text corresponding to the target needs.
[0135] Among them, the clinical scientific research model is obtained through training and optimization based on the scientific research database and the scientific research corpus. For relevant explanations, please refer to the explanations about the clinical scientific research model in the above embodiment, which will not be elaborated here.
[0136] Step 3: The clinical research model determines the first clinical research information according to the answer type corresponding to the target question text, or determines the second clinical research information by calling the medical big data platform through the target interface.
[0137] Among them, if the answer type of the target question text is a preset type, the first clinical research information is directly determined according to the target question text, otherwise the second clinical research information is determined by calling the medical big data platform according to the target question text through the target interface.
[0138] It should be understood that the clinical research model calls the data preprocessing function of the medical big data platform through the target interface, and can also provide data preprocessing methods for a variety of statistical software, and generate standard data sets that meet scientific research needs; call the data retrieval function to retrieve and save the research population; call the statistical analysis function and use the platform's existing statistical models to model and analyze the research data.
[0139] In the embodiments of the present application, various forms of system docking can be supported through different interfaces, not limited to the medical big data platform, such as the hospital business system, the in-hospital PC terminal, and the personalized configuration function module of the mobile terminal. Among them, the mobile terminal only supports literature retrieval, article review, and paper polishing functions.
[0140] It should be understood that if the answer to the target question text involves querying / analyzing real-world data (clinical data), in order to protect patient privacy and data security, the clinical research model can call the data query and analysis functions of the medical database or medical big data platform through the target interface, and the medical database or medical big data platform returns the clinical research information corresponding to the target demand. For example, the clinical research model can determine whether the target question text involves real clinical data, and if so, give an instruction to call the target interface, and the medical database or medical big data platform calls the corresponding function through the instruction, such as querying or retrieving the research population, cleaning and statistically analyzing the research data, etc.
[0141] Step 4: Determine the research steps according to the clinical research process module and execute the corresponding number of cycles.
[0142] It should be understood that the clinical research process module can include multiple research links, which are connected in series, and different target requirements may be targeted at different research links. The clinical research module can cover most medical research needs, meet the needs of clinical research, help researchers clarify their research ideas, and make clinical research work more standardized, automated, and streamlined.
[0143] As an example, assuming that the target requirement indicates the first research phase, the clinical research information can be directly output according to the result of step 3.
[0144] As another example, assuming that the target requirement indicates the second research link, then after the target requirement is input, the target requirement needs to indicate the first research link, and based on the first research link, steps 1 to 3 are executed to obtain the clinical research information of the first research link (referred to as the first information), and then the first information is input into the clinical research model again. At this time, the target requirement indicates the second research link, and steps 1 to 3 are executed based on the second research link and the first information to obtain the clinical research information of the second research link. That is, if the target requirement does not indicate the first research link, the n-1 links before the nth research link (where n is an integer greater than 1) need to execute the above steps 1 to 3 in sequence, and use the results of each execution as the input of the next cycle, that is, the input of the nth research link is the target requirement indicating the nth link and the n-1th information.
[0145] Step 5: After obtaining the clinical research information output by the clinical research model, add the clinical research information and the corresponding target question text as dialogue text to the research corpus.
[0146] It should be understood that storing clinical research information and the corresponding target question text (i.e., dialogue text) in a scientific research corpus can enhance the data quality and richness of the scientific research corpus. Further, the updated scientific research corpus and scientific research database can be used to train and optimize the clinical research model, thereby improving the accuracy and reliability of the clinical research model.
[0147] Based on the clinical research information acquisition method provided in the above embodiment, the present application embodiment also provides a clinical research information acquisition device, see Figure 6 , which is a schematic diagram of the structure of a clinical research information acquisition device provided in an embodiment of the present application, combined with Figure 6 As shown, the clinical research information acquisition device 600 provided in the embodiment of the present application may include:
[0148] The question text determination module 601 is used to determine the target question text according to the target requirements through the pre-trained clinical research model; the target question text indicates the questions related to the target research link; the clinical research model is trained based on the research database and the research corpus, the research database includes a plurality of annotated medical data, and the research corpus includes question and answer corpora corresponding to a plurality of research links;
[0149] A determination module 602 is used to determine whether the answer type corresponding to the target question text is a preset type;
[0150] The first information acquisition module 603 is used to determine the first clinical research information according to the target question text through the clinical research model if the answer type is a preset type; the type of the first clinical research information is a preset type;
[0151] The second information acquisition module 604 is used to determine the target interface according to the answer type if the answer type is not a preset type, and determine the second clinical research information according to the target question text through the target interface, and the type of the second clinical research information is not a preset type.
[0152] In a possible implementation, when the target research link is the second research link, the first information acquisition module 603 is used to determine the first clinical research information according to the historical clinical research information and the question text through the clinical research model; the historical clinical research information is the clinical research information corresponding to the first research link; the first research link is associated with the second research link;
[0153] Correspondingly, the second information acquisition module 604 is used to determine the second clinical research information according to the target question text and the historical clinical research information through the target interface.
[0154] In a possible implementation, before the first information acquisition module 603, it further includes: a cutting module, which is used to cut the clinical research information corresponding to the first research link to obtain historical clinical research information.
[0155] In a possible implementation, the first information acquisition module 603 is used to obtain input data corresponding to the target research link; through the clinical research model, the first clinical research information is determined according to the target problem text and the input data.
[0156] In one possible implementation, a scientific research corpus is obtained in the following manner: a scientific research corpus construction module is used to obtain a set of clinical research questions; the set of clinical research questions includes multiple clinical research questions; each clinical research question in the set of clinical research questions is classified and labeled to obtain multiple subsets; the type and labeling results of each clinical research question in the subset are the same; the question template corresponding to each subset is determined; the question template is obtained based on the clinical research question in the corresponding subset; and a scientific research corpus is established based on multiple question and answer templates.
[0157] In a possible implementation, the second information acquisition module 604 is used to call a target tool or a target database through a target interface; and process the target question text according to the target tool or the target database to obtain the second clinical research information.
[0158] In a possible implementation, the device 600 may further include: a training module, which is used to train the clinical research model using the clinical research information after determining the clinical research information; the clinical research information is the first clinical research information or the second clinical research information.
[0159] The clinical research information acquisition device provided in the embodiment of the present application has the same beneficial effects as the clinical research information acquisition method provided in the above embodiment, so it will not be repeated here.
[0160] The present application also provides a computer device, such as Figure 7 As shown, the computer device 01 is in the form of a general computing device. The components of the computer device 01 may include but are not limited to: one or more processors or processing units 03, a system memory 08, and a bus 04 connecting different system components (including the system memory 08 and the processing unit 03).
[0161] Bus 04 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor or a local bus using any of a variety of bus structures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnect (PCI) bus.
[0162] The computer device 01 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer device 01, including volatile and non-volatile media, removable and non-removable media.
[0163] The system memory 08 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 09 and / or cache memory 10. The computer device 01 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 11 may be used to read and write non-removable, non-volatile magnetic media ( Figure 3 not shown, usually called a "hard drive"). Although Figure 3 Not shown in the figure, a disk drive for reading and writing a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 04 via one or more data medium interfaces. The memory 08 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.
[0164] A program / utility 12 having a set (at least one) of program modules 13 may be stored, for example, in the memory 08, such program modules 13 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination thereof may include an implementation of a network environment. The program modules 13 generally perform the functions and / or methods of the embodiments described herein.
[0165] The computer device 01 may also communicate with one or more external devices 02 (e.g., keyboard, pointing device, display 07, etc.), one or more devices that enable a user to interact with the computer device 01, and / or any device that enables the computer device 01 to communicate with one or more other computing devices (e.g., network card, modem, etc.). Such communication may be performed via input / output (I / O) interface 06. Furthermore, the computer device 01 may also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN) and / or public network, such as the Internet) via network adapter 05. Figure 3 As shown, the network adapter 05 communicates with other modules of the computer device 01 via the bus 04. It should be understood that although Figure 3 Not shown, other hardware and / or software modules may be used in conjunction with computer device 01, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0166] The processor unit 03 executes various functional applications and data processing by running the programs stored in the system memory 08, such as implementing the clinical research information acquisition method provided in any embodiment of the present application.
[0167] The embodiments of the present application also provide a corresponding computer storage medium, in which codes are stored. When the codes are executed, the device executing the codes implements the clinical research information acquisition method described in any embodiment of the present application.
[0168] It should be noted that each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and equipment embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments. The device and equipment embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components indicated as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. A person of ordinary skill in the art can understand and implement it without paying creative labor.
[0169] The "first" and "second" in the names such as "first" and "second" (if any) mentioned in the embodiments of the present application are only used as name identifiers and do not represent the first or second in order.
[0170] Through the description of the above implementation methods, it can be known that those skilled in the art can clearly understand that all or part of the steps in the above-mentioned embodiment method can be implemented by means of software plus a general hardware platform. Based on such an understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a storage medium, such as a read-only memory (ROM) / RAM, a magnetic disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network communication device such as a router) to execute the methods described in each embodiment of the present application or some parts of the embodiments.
[0171] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A method for obtaining clinical research information, characterized in that: The method comprises: Obtaining target requirements of the target subject; the target requirements indicate target research links of the clinical research; The target question text is determined according to the target demand through the pre-trained clinical research model; the target question text indicates the problem related to the target research link; the clinical research model is trained based on the research database and the research corpus, the research database includes a plurality of annotated medical data, and the research corpus includes question and answer corpora corresponding to a plurality of research links; Determine whether the answer type corresponding to the target question text is a preset type; If the answer type is a preset type, determining the first clinical research information according to the target question text through the clinical research model; the type of the first clinical research information is the preset type; If the answer type is not the preset type, a target interface is determined according to the answer type, and second clinical research information is determined according to the target question text through the target interface, and the type of the second clinical research information is the answer type.
2. The method according to claim 1, characterized in that When the target research link is the second research link, determining the first clinical research information according to the target problem text through the clinical research model includes: Determine the first clinical research information according to the historical clinical research information and the question text through the clinical research model; the historical clinical research information is the clinical research information corresponding to the first research link; the first research link is associated with the second research link; Correspondingly, determining the second clinical research information according to the target question text through the target interface includes: determining the second clinical research information according to the target question text and the historical clinical research information through the target interface.
3. The method according to claim 2, characterized in that Before determining the first clinical research information according to the historical clinical research information and the question text through the clinical research model, the method further includes: The clinical research information corresponding to the first research link is cut and processed to obtain historical clinical research information.
4. The method according to claim 1, characterized in that Determining the first clinical research information according to the target question text through the clinical research model includes: Obtaining input data corresponding to the target research link; The first clinical research information is determined according to the target question text and the input data through the clinical research model.
5. The method according to claim 1, characterized in that The scientific research corpus is obtained by: Obtaining a set of clinical research questions; the set of clinical research questions includes multiple clinical research questions; Classifying and labeling each clinical research question in the clinical research question set to obtain multiple subsets; the types and labeling results of each clinical research question in the subsets are the same; Determine the question template corresponding to each of the subsets; The question template is obtained based on the clinical research questions in the corresponding subset; Establish a scientific research corpus based on multiple question-answer templates.
6. The method according to claim 1, characterized in that Determining the second clinical research information according to the target question text through the target interface includes: Calling a target tool or a target database through the target interface; The target question text is processed according to the target tool or the target database to obtain second clinical research information.
7. The method according to claim 1, characterized in that Also includes: After the clinical research information is determined, the clinical research model is trained using the clinical research information; the clinical research information is the first clinical research information or the second clinical research information.
8. A clinical research information acquisition device, characterized in that: The device comprises: An acquisition module is used to acquire target requirements of a target subject; the target requirements indicate target research links of a clinical study; A question text determination module is used to determine a target question text according to the target demand through a pre-trained clinical research model; the target question text indicates a question related to the target research link; the clinical research model is trained based on a research database and a research corpus, the research database includes a plurality of annotated medical data, and the research corpus includes question and answer corpora corresponding to a plurality of research links; A judgment module, used to judge whether the answer type corresponding to the target question text is a preset type; A first information acquisition module, configured to determine first clinical research information according to the target question text through the clinical research model if the answer type is a preset type; the type of the first clinical research information is the preset type; The second information acquisition module is used to determine the target interface according to the answer type if the answer type is not the preset type, and determine the second clinical research information according to the target question text through the target interface, and the type of the second clinical research information is not the preset type.
9. A computer device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for acquiring clinical research information as described in any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the clinical research information acquisition method according to any one of claims 1 to 7.