Chinese patent medicine specification exceeding evidence-based evaluation system and method based on large language model
Through the evidence-based evaluation system for the Chinese patent medicines super-instructions based on the large language model, the problem of lack of evidence-based support in the use of the Chinese patent medicines super-instructions is solved, efficient and accurate literature processing and evaluation is achieved, reliable clinical decision-making support is provided, and the development of the traditional Chinese medicine industry is promoted.
Patent Information
- Application Number
- CN202510440805.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-09
- Publication Date
- 2025-07-29
AI Technical Summary
The use of super-instructions for traditional Chinese patent medicines in the prior art lacks systematic evidence-based support, resulting in high drug use risks and low manual search and analysis efficiency, making it difficult to process large-scale literature information, and reducing safety and effectiveness.
An evidence-based evaluation system for Chinese patent medicines based on large language models is adopted, including data collection, literature classification, key information extraction, automatic grading of evidence and intelligent evidence-based report generation model to achieve automated processing and evaluation.
It has achieved scientific and objective evaluation of the super-instructions of Chinese patent medicines, improved the efficiency and accuracy of literature screening and information extraction, provided a reliable basis for clinical decision-making, and promoted the healthy development of the traditional Chinese medicine industry.
Smart Images
Figure CN120386866A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the cross - technical field of traditional Chinese medicine and artificial intelligence, and particularly to a system and method for evidence - based evaluation of Chinese patent medicine beyond the scope of the approved label based on large language models. Background Art
[0002] With the wide application of Chinese patent medicine in clinical practice, the phenomenon of its use beyond the scope of the approved label has become increasingly common, involving indications, dosages, usage methods or populations not explicitly approved by the label. Although some have actual curative effects, due to the lack of systematic evidence - based support, there are medication risks and it is difficult to obtain regulatory approval.
[0003] The current evidence - based evaluation methods mainly rely on manual retrieval and analysis, which are inefficient, highly subjective, difficult to process large - scale literature information, reduce the safety and effectiveness of the use of Chinese patent medicine beyond the scope of the approved label, and cannot promote the healthy development of traditional Chinese medicine.
[0004] Therefore, we propose a system and method for evidence - based evaluation of Chinese patent medicine beyond the scope of the approved label based on large language models. Summary of the Invention
[0005] The present invention mainly solves the technical problems existing in the above - mentioned prior art, and provides a system and method for evidence - based evaluation of Chinese patent medicine beyond the scope of the approved label based on large language models.
[0006] To achieve the above object, the present invention adopts the following technical solutions. The evidence - based evaluation system of Chinese patent medicine beyond the scope of the approved label based on large language models includes a data collection model, a literature classification model, a key information extraction model, an evidence automatic grading model, and an intelligent evidence - based report generation model. The data collection model includes a data source interface, a data pre - processing module, and a quality control module.
[0007] Preferably, the literature classification model includes a deep learning algorithm, a feature extraction module, and a classification decision module, which can automatically classify the literature related to Chinese patent medicine and improve the efficiency of literature screening.
[0008] Preferably, the key information extraction model includes a key information extraction template, a Prompt template, and a large language model, which can accurately extract the key information related to the use of Chinese patent medicine beyond the scope of the approved label from the classified literature.
[0009] Preferably, the evidence automatic grading model includes an evidence grading rule library, an evidence evaluation module, and a grading decision module, which can automatically grade the extracted key information according to the preset evidence grading criteria to ensure the accuracy and reliability of the evaluation results.
[0010] Preferably, the intelligent evidence-based report generation model includes a report template library, a data integration module, and a report generation module, which can automatically generate a well-structured and detailed evidence-based evaluation report based on the classified key information, providing strong support for clinical decision-making.
[0011] The method for evidence-based evaluation of off-label use of Chinese patent medicines based on large language models includes the above-mentioned evidence-based evaluation system for off-label use of Chinese patent medicines based on large language models, and specifically includes the following steps:
[0012] Step 1: Construct a data collection model: Obtain various types of literature, clinical trial data, etc. related to Chinese patent medicines through a data source interface. The data preprocessing module processes the data, such as cleaning and format conversion, and the quality control module monitors and evaluates the data quality to ensure the accuracy and integrity of the data;
[0013] Step 2: Construct a literature classification model: First, based on the Classification Table of Clinical Evidence Types of Chinese Patent Medicines, train a large language model through supervised fine-tuning, and then use an iteratively fine-tuned literature type dataset with manual annotation to optimize the model classification performance. Finally, output the classification results to improve the literature processing efficiency;
[0014] Step 3: Construct a key information extraction model: First, construct an information extraction template for Chinese patent medicine literature, and then combine a public dataset and manually annotated data to fine-tune the large language model to achieve structured information extraction. Finally, standardize the extraction results through ontology term matching;
[0015] Step 4: Construct an evidence automatic grading model: First, based on the GRADE system or the evidence-based grading standard of traditional Chinese medicine, construct a training set, and then automatically grade the literature evidence through a large language model. Introduce a manual verification mechanism to correct biases. Finally, output the objective evidence level to reduce the influence of human subjectivity;
[0016] Step 5: Construct an intelligent evidence-based report generation model: First, integrate the classification, extraction, and grading results, combine the Chinese patent medicine instruction dataset, then design an algorithm to automatically identify off-label indications, count the evidence frequency, and finally generate a structured report through a large language model, including a recommendation level, an evidence summary, and references.
[0017] The present invention provides an evidence-based evaluation system and method for off-label use of Chinese patent medicines based on large language models.
[0018] It has the following beneficial effects:
[0019] 1. The evidence-based evaluation system and method for off-label use of Chinese patent medicines based on large language models integrates a data collection model, a literature classification model, a key information extraction model, an evidence automatic grading model, and an intelligent evidence-based report generation model, achieving a scientific and objective evaluation of the off-label use of Chinese patent medicines, providing strong support for clinical decision-making, being able to comprehensively and efficiently process Chinese patent medicine-related literature and clinical trial data, extract key information, conduct evidence grading, and automatically generate evidence-based evaluation reports, which will help improve the safety and effectiveness of the off-label use of Chinese patent medicines and promote the healthy development of traditional Chinese medicine.
[0020] 2. The evidence-based evaluation system and method for off-label use of Chinese patent medicines based on large language models can accurately classify Chinese patent medicine-related literature automatically by setting up a literature classification model, avoiding the subjectivity and inefficiency of manual classification, improving the accuracy and efficiency of literature screening, and laying a solid foundation for subsequent information extraction and evidence grading.
[0021] 3. The evidence-based evaluation system and method for off-label use of Chinese patent medicines based on large language models can quickly and accurately extract key information directly related to the off-label use of Chinese patent medicines, such as indications, dosage and administration, and efficacy evaluation, from a large number of classified literatures by setting up a key information extraction model, effectively avoiding information omission and misunderstanding, and ensuring the rigor and scientificity of the evaluation process.
[0022] 4. The evidence-based evaluation system and method for off-label use of Chinese patent medicines based on large language models can objectively and fairly grade the extracted key information automatically according to the internationally recognized GRADE system or the evidence-based grading criteria unique to traditional Chinese medicine by setting up an evidence automatic grading model, avoiding the subjectivity and inconsistency of manual grading, ensuring the accuracy and authority of the evaluation results, and providing a more reliable basis for clinical decision-making.
[0023] 5. The evidence-based evaluation system and method for off-label use of Chinese patent medicines based on large language models can quickly generate a detailed and clearly structured evidence-based evaluation report according to the graded key information by setting up an intelligent evidence-based report generation model. This report not only includes key contents such as recommendation grades and evidence summaries, but also lists the references in detail, providing a comprehensive and objective evaluation basis for clinical doctors and researchers on the off-label use of Chinese patent medicines. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 is the system architecture diagram of the present invention;
[0025] Figure 2 is the method flow diagram of the present invention. DETAILED DESCRIPTION OF THE INVENTION
[0026] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary. For those of ordinary skill in the art, without creative efforts, other implementation drawings can also be obtained based on the provided drawings.
[0027] The structures, proportions, sizes, etc. illustrated in this specification are only used to cooperate with the content disclosed in the specification for those familiar with this technology to understand and read, and are not used to limit the limited conditions under which the present invention can be implemented. Therefore, they do not have technical substantive significance. Any modification of the structure, change in the proportional relationship, or adjustment of the size, without affecting the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope covered by the technical content disclosed by the present invention.
[0028] It should be noted that similar reference numerals and letters indicate similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.
[0029] In the description of the embodiments of the present invention, it should be noted that the orientation or positional relationship indicated by terms such as "center", "upper", "lower", "inner", "outer", "side", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of this invention is usually placed during use. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation to the present invention. In addition, terms such as "first" and "second" are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.
[0030] In the description of the embodiments of the present invention, it should also be noted that unless otherwise clearly specified and limited, the terms "set", "installed", "connected", and "coupled" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present invention can be understood according to specific situations.
[0031] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope protected by the present invention.
[0032] Example 1: A traditional Chinese medicine over-instruction evidence-based evaluation system based on a large language model, as Figure 1 shown, includes a data collection model, a literature classification model, a key information extraction model, an evidence automatic grading model, and an intelligent evidence-based report generation model. The data collection model includes a data source interface, a data preprocessing module, and a quality control module. The literature classification model includes a deep learning algorithm, a feature extraction module, and a classification decision module, which can automatically classify the literature related to traditional Chinese medicine and improve the efficiency of literature screening. The key information extraction model includes a key information extraction template, a Prompt template, and a large language model, which can accurately extract the key information related to the off-label use of traditional Chinese medicine from the classified literature. The evidence automatic grading model includes an evidence grading rule library, an evidence evaluation module, and a grading decision module, which can automatically grade the extracted key information according to the preset evidence grading criteria to ensure the accuracy and reliability of the evaluation results. The intelligent evidence-based report generation model includes a report template library, a data integration module, and a report generation module, which can automatically generate a well-structured and detailed evidence-based evaluation report based on the graded key information to provide strong support for clinical decision-making. By integrating the data collection model, the literature classification model, the key information extraction model, the evidence automatic grading model, and the intelligent evidence-based report generation model, a scientific and objective evaluation of the off-label use of traditional Chinese medicine is realized, providing strong support for clinical decision-making, being able to comprehensively and efficiently process the literature and clinical trial data related to traditional Chinese medicine, extract key information, conduct evidence grading, and automatically generate an evidence-based evaluation report, which will help improve the safety and effectiveness of the off-label use of traditional Chinese medicine and promote the healthy development of the traditional Chinese medicine cause.
[0033] Example 2: On the basis of Example 1, as Figure 1As shown in the figure, the data collection model includes a data source interface, a data preprocessing module, and a quality control module. The literature classification model includes a deep learning algorithm, a feature extraction module, and a classification decision module, which can automatically classify the literature related to Chinese patent medicines and improve the efficiency of literature screening. The key information extraction model includes a key information extraction template, a Prompt template, and a large language model, which can accurately extract the key information related to off-label use of Chinese patent medicines from the classified literature. The evidence automatic grading model includes an evidence grading rule base, an evidence evaluation module, and a grading decision module, which can automatically grade the extracted key information according to the preset evidence grading criteria to ensure the accuracy and reliability of the evaluation results. The intelligent evidence-based report generation model includes a report template library, a data integration module, and a report generation module, which can automatically generate a well-structured and detailed evidence-based evaluation report based on the graded key information to provide strong support for clinical decision-making. By setting up the literature classification model, it is possible to automatically and accurately classify the literature related to Chinese patent medicines, avoiding the subjectivity and inefficiency of manual classification, improving the accuracy and efficiency of literature screening, and laying a solid foundation for subsequent information extraction and evidence grading.
[0034] Example 3: On the basis of Example 1 and Example 2, as Figure 1 shown in the figure, the literature classification model includes a deep learning algorithm, a feature extraction module, and a classification decision module, which can automatically classify the literature related to Chinese patent medicines and improve the efficiency of literature screening. The key information extraction model includes a key information extraction template, a Prompt template, and a large language model, which can accurately extract the key information related to off-label use of Chinese patent medicines from the classified literature. The evidence automatic grading model includes an evidence grading rule base, an evidence evaluation module, and a grading decision module, which can automatically grade the extracted key information according to the preset evidence grading criteria to ensure the accuracy and reliability of the evaluation results. The intelligent evidence-based report generation model includes a report template library, a data integration module, and a report generation module, which can automatically generate a well-structured and detailed evidence-based evaluation report based on the graded key information to provide strong support for clinical decision-making. By setting up the key information extraction model, it is possible to quickly and accurately extract the key information directly related to off-label use of Chinese patent medicines, such as indications, dosage and usage, and efficacy evaluation, from a large number of classified literature, effectively avoiding information omission and misunderstanding, and ensuring the rigor and scientificity of the evaluation process.
[0035] Example 4: On the basis of Example 1, Example 2, and Example 3, as Figure 1As shown, the key information extraction model includes a key information extraction template, a Prompt template, and a large language model, and can accurately extract key information related to the off-label use of Chinese patent medicines from the classified literature. The evidence automatic grading model includes an evidence grading rule base, an evidence evaluation module, and a grading decision module, and can automatically grade the extracted key information according to the preset evidence grading criteria to ensure the accuracy and reliability of the evaluation results. The intelligent evidence-based report generation model includes a report template library, a data integration module, and a report generation module, and can automatically generate an evidence-based evaluation report with clear structure and detailed content according to the graded key information to provide strong support for clinical decision-making. By setting up an evidence automatic grading model, it is possible to objectively and fairly automatically grade the extracted key information according to the internationally recognized GRADE system or the evidence-based grading criteria unique to traditional Chinese medicine, avoiding the subjectivity and inconsistency of manual grading, ensuring the accuracy and authority of the evaluation results, and providing a more reliable basis for clinical decision-making.
[0036] Example 5: On the basis of Example 1, Example 2, Example 3, and Example 4, as Figure 2As shown, the method for evidence-based evaluation of off-label use of Chinese patent medicines based on large language models includes the above-mentioned evidence-based evaluation system for off-label use of Chinese patent medicines based on large language models, and specifically includes the following steps: The first step: Construct a data collection model: Obtain various types of literature, clinical trial data, etc. related to Chinese patent medicines through the data source interface. The data preprocessing module processes the data such as cleaning and format conversion, and the quality control module monitors and evaluates the data quality to ensure the accuracy and integrity of the data; The second step: Construct a literature classification model: First, based on the "Classification Table of Clinical Evidence Types of Chinese Patent Medicines", train the large language model through supervised fine-tuning, and then use the manually labeled literature type data set for iterative fine-tuning to optimize the model classification performance. Finally, output the classification results to improve the literature processing efficiency; The third step: Construct a key information extraction model: First, construct an information extraction template for Chinese patent medicine literature, and then combine the public data set with the manually labeled data to fine-tune the large language model to achieve structured information extraction. Finally, through ontology term matching, standardize the extraction results; The fourth step: Construct an evidence automatic grading model: First, based on the GRADE system or the evidence-based grading standard of traditional Chinese medicine, construct a training set, and then automatically grade the literature evidence through the large language model, and introduce a manual verification mechanism to correct the deviation. Finally, output the objective evidence level to reduce the influence of human subjectivity; The fifth step: Construct an intelligent evidence-based report generation model: First, integrate the classification, extraction and grading results, combine the Chinese patent medicine instruction manual data set, and then design an algorithm to automatically identify off-label indications, count the evidence frequency, and finally generate a structured report through the large language model, including the recommendation level, evidence summary and references. By setting up an intelligent evidence-based report generation model, an evidence-based evaluation report with detailed content and clear structure can be quickly generated according to the graded key information. This report not only includes key contents such as the recommendation level and evidence summary, but also lists the references in detail, providing a comprehensive and objective basis for the evaluation of off-label use of Chinese patent medicines for clinicians and researchers.
[0037] Working principle of the present invention: First, through the data collection model, various types of literature and clinical trial data related to traditional Chinese medicine are comprehensively collected to ensure the comprehensiveness and diversity of information. Subsequently, the literature classification model automatically classifies these literatures accurately, effectively improving the efficiency of literature screening and laying a solid foundation for subsequent steps. The key information extraction model further quickly and accurately extracts key information directly related to the off-label use of traditional Chinese medicine from the classified literatures, such as indications, dosage and usage, efficacy evaluation, etc., ensuring the integrity and accuracy of information. The evidence automatic grading model objectively and fairly automatically grades the extracted key information according to the internationally recognized GRADE system or the evidence-based grading criteria unique to traditional Chinese medicine, further improving the accuracy and authority of the evaluation results. Finally, the intelligent evidence-based report generation model automatically generates an evidence-based evaluation report with clear structure and detailed content according to the graded key information, providing a comprehensive and objective basis for the evaluation of the off-label use of traditional Chinese medicine for clinicians and researchers. The entire system and method achieve a scientific and objective evaluation of the off-label use of traditional Chinese medicine, provide strong support for clinical decision-making, will help improve the safety and effectiveness of the off-label use of traditional Chinese medicine, and promote the healthy development of the traditional Chinese medicine cause.
[0038] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.
Claims
1. A traditional Chinese medicine off-label evidence-based evaluation system based on large language models, characterized in that, It includes a data collection model, a literature classification model, a key information extraction model, an evidence automatic grading model, and an intelligent evidence-based report generation model. The data collection model includes a data source interface, a data preprocessing module, and a quality control module.
2. The evidence-based evaluation system for traditional Chinese patent medicine beyond the scope of approved labeling based on large language models according to claim 1, wherein: The literature classification model includes a deep learning algorithm, a feature extraction module, and a classification decision module.
3. The evidence-based evaluation system for off-label use of traditional Chinese patent medicines based on large language models according to claim 1, characterized in that: The key information extraction model includes a key information extraction template, a Prompt template, and a large language model.
4. The evidence-based evaluation system for off-label use of Chinese patent medicines based on large language models according to claim 1, wherein: The evidence automatic grading model includes an evidence grading rule library, an evidence evaluation module, and a grading decision module.
5. The evidence-based evaluation system for off-label use of traditional Chinese patent medicines based on large language models according to claim 1, characterized in that: The intelligent evidence-based report generation model includes a report template library, a data integration module, and a report generation module.
6. The evidence-based evaluation method for the off-label use of traditional Chinese patent medicines based on large language models, characterized in that, It includes the Chinese patent medicine off-label evidence-based evaluation system based on the large language model described in any one of claims 1-5, and specifically includes the following steps: The first step: Construct a data collection model: Obtain various types of literature, clinical trial data, etc. related to Chinese patent medicine through the data source interface. The data preprocessing module processes the data such as cleaning and format conversion, and the quality control module monitors and evaluates the data quality to ensure the accuracy and integrity of the data; The second step: Construct a literature classification model: First, based on the "Classification Table of Clinical Evidence Types of Chinese Patent Medicine", train the large language model through supervised fine-tuning, then use the manually labeled literature type data set for iterative fine-tuning to optimize the model classification performance, and finally output the classification result to improve the literature processing efficiency; The third step: Construct a key information extraction model: First, construct a Chinese patent medicine literature information extraction template, then combine the public data set with the manually labeled data to fine-tune the large language model to achieve structured information extraction, and finally standardize the extraction result through ontology term matching; The fourth step: Construct an evidence automatic grading model: First, based on the GRADE system or the evidence-based grading standard of traditional Chinese medicine, construct a training set, then automatically grade the literature evidence through the large language model, and introduce a manual verification mechanism to correct the deviation, and finally output the objective evidence level to reduce the influence of human subjectivity; The fifth step: Construct an intelligent evidence-based report generation model: First, integrate the classification, extraction, and grading results, combine the Chinese patent medicine instruction data set, then design an algorithm to automatically identify off-label indications, count the evidence frequency, and finally generate a structured report through the large language model, including the recommendation level, evidence summary, and reference literature.