Clinical test data analysis system, method and equipment based on large model and medium

Abstract

The invention discloses a clinical test data analysis system based on a large model, which comprises a data acquisition module, a preprocessing module, a training module, an analysis module and an output module, data acquisition is efficient and supports multiple modes, preprocessing adopts an intelligent algorithm, model training is advanced and has efficient indexes, analysis covers multiple aspects and has evaluation indexes, and result output is diversified and supports customization. The invention further comprises a corresponding data analysis method, related equipment and a medium, and each link is provided with an innovative technology and an evaluation mode to guarantee the performance and the quality; the system realizes efficient data acquisition and integration, intelligent processing and quality improvement, constructs a powerful analysis model, provides various visual result display and interpretation, has real-time updating and personalized service capabilities, can accurately evaluate the curative effect and safety, provides a scientific basis for clinical decisions, assists in improving the efficiency and accuracy of medical data analysis, and has a wide application prospect. Different user requirements are met, and development of the medical industry is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of clinical trial data analysis systems, and more specifically, to a clinical trial data analysis system, method, device and medium based on a large model. Background Art

[0002] With the continuous development of medical technology and the increasing number of clinical trials, efficient analysis and accurate interpretation of clinical trial data have become crucial. However, traditional clinical trial data analysis methods face many challenges and difficulties.

[0003] In terms of data collection, traditional methods often rely on manual extraction of data from various scattered data sources, which is not only inefficient, but also prone to data omissions, errors, and inconsistent formats. Different medical systems and devices use different data formats and storage methods, making data integration difficult. For example, the data structure in the electronic medical record system and the data in the laboratory information management system are difficult to be directly compatible, requiring a lot of manual conversion and sorting work.

[0004] In the data processing stage, traditional data analysis methods usually use simple statistical tools and fixed analysis processes, which cannot cope with the complexity and diversity of clinical trial data. For large amounts of high-dimensional data, such as data containing patients' genetic information, physiological indicators, treatment processes, and other aspects, traditional methods find it difficult to deeply explore potential associations and patterns. Moreover, the data cleaning process mainly relies on manual screening and removal of outliers, lacking intelligent algorithms and tools, making it difficult to ensure data quality.

[0005] In terms of model building and analysis, traditional statistical models and machine learning algorithms have limited capabilities when processing large-scale clinical trial data. They have difficulty capturing complex nonlinear relationships and dynamic change characteristics in the data, and cannot meet the needs of modern clinical trials for accurate analysis and prediction. For example, in drug efficacy evaluation and adverse reaction prediction, traditional methods have low accuracy and reliability and cannot provide strong support for clinical decision-making.

[0006] In addition, traditional data analysis results are presented in a single way, often in simple tables and statistical charts, which are difficult to intuitively reflect the connotation and trend of the data. This makes it difficult for clinical researchers and doctors to quickly understand and apply the analysis results, affecting the efficiency and quality of decision-making.

[0007] In summary, the existing clinical trial data analysis technology can no longer meet the needs of the rapid development of the medical industry. An innovative large-model-based data analysis system, method, equipment and medium are urgently needed to solve these problems. Summary of the invention

[0008] The clinical trial data analysis system, method, device and medium based on a large model proposed in the present invention are used to solve the problems mentioned in the above-mentioned prior art.

[0009] In order to achieve the above-mentioned object, the present invention adopts the following technical scheme: a clinical trial data analysis system based on a large model, comprising a data acquisition module, a data preprocessing module, a model training module, a data analysis module and a result output module;

[0010] The data collection module is used to efficiently collect raw data from various clinical trial data sources, including electronic medical record systems, laboratory information management systems, medical equipment data, etc. It uses advanced data interface technology and secure data encryption transmission methods to ensure data accuracy and completeness; for example, it supports automatic conversion and parsing of multiple data formats. Data collection efficiency evaluation indicators Where E collect represents the data collection efficiency, N collected Indicates the amount of data collected per unit time, T collect Indicates the collection time;

[0011] The data preprocessing module cleans, standardizes, and annotates the collected data. It uses intelligent data cleaning algorithms to automatically identify and remove outliers, duplicate data, and erroneous data. It uses standardized data format conversion methods to unify data formats and accurately annotate data based on professional medical knowledge and annotation rules. The data cleaning effect evaluation formula is: Where Q represents the quality improvement rate after data cleaning, N clean Indicates the amount of data after cleaning, N error Indicates the amount of error data before cleaning, N original Indicates the amount of raw data;

[0012] The model training module uses large-scale clinical trial data to train large models, and uses advanced deep learning algorithms and distributed training technologies to build high-performance data analysis models, such as neural network models based on the Transformer architecture. It optimizes model parameters and structures through large-scale data iterative training. The model training convergence index C = |f(x n )-f(x n-1 )|<∈, where C represents the model convergence condition, f(x n ) and f(x n-1 ) represent the model output values ​​of two adjacent iterations, ∈ is the preset minimum value;

[0013] The data analysis module uses the trained large model to conduct in-depth analysis of the pre-processed data, including efficacy evaluation, safety analysis, and patient characteristics analysis. It mines the potential information of the data through the powerful computing power and intelligent algorithm of the model, providing a scientific basis for clinical trial decision-making and optimization.efficacy =α×P recovery +β×T improve Among them, R efficacy represents the efficacy evaluation value, α and β are weight coefficients, P recovery Represents the patient recovery rate and other related indicators, T improve Indicates relevant indicators such as the time it takes for patients’ symptoms to improve;

[0014] The result output module displays the analysis results to users in an intuitive and easy-to-understand form, supports a variety of data visualization methods such as charts, reports, dynamic images, etc., and provides detailed data reports and analysis interpretations. Where S represents the user's satisfaction with the result display, N satisfied The number of satisfied users, N total Indicates the total number of users.

[0015] Preferably, the data interface technology in the data acquisition module supports two modes: real-time data acquisition and batch data acquisition, and can flexibly select the acquisition method according to different application scenarios and data sources. At the same time, data encryption transmission uses the Advanced Encryption Standard AES algorithm to ensure data security.

[0016] Preferably, the distributed training technology in the model training module adopts a combination of data parallelism and model parallelism. Large-scale clinical trial data is divided into multiple small data blocks and distributed to multiple computing nodes for parallel training. At the same time, different parts of the model are distributed to different nodes for collaborative training to improve training efficiency. The training efficiency improvement index is: Where T represents the percentage of training efficiency improvement, T single Represents the single node training time, T distributed Represents the distributed training time.

[0017] Preferably, the safety analysis in the data analysis module adopts a multi-dimensional data analysis method. A comprehensive evaluation model is constructed by combining various clinical trial indicators and patient baseline characteristics, for example, analyzing the relationship between factors such as drug dosage, treatment time, patient age, gender, and safety. Safety evaluation index S safety =γ×A adverse +δ×F frequency , where S safety represents the safety assessment value, γ and δ are weight coefficients, A adverse Indicates the probability of adverse events and other related indicators, F frequency Indicates relevant indicators such as the frequency of adverse events.

[0018] Preferably, the result output module supports personalized result customization and push service, and users can select specific analysis results and data display methods according to their own needs, and set the time and method of automatic push.

[0019] Preferably, a clinical trial data analysis method based on a large model comprises the following steps:

[0020] S1. Data collection: Use the data collection module to collect raw data from various clinical trial data sources, and conduct preliminary sorting and classification, and record the data source, time and related clinical trial information;

[0021] S2, data preprocessing: the collected data is transferred to the data preprocessing module for cleaning, standardization, labeling and other processing operations;

[0022] S3. Model training: Input the preprocessed clinical trial data into the model training module, use advanced deep learning algorithms and distributed training technology to train large models, and select appropriate model architecture and parameter settings based on data characteristics and analysis requirements;

[0023] S4. Data analysis: Use the trained large model to conduct in-depth analysis of clinical trial data, and select corresponding analysis algorithms and model parameters based on user needs and set analysis goals;

[0024] S5. Result output: The analysis results are displayed and output through the result output module, and the appropriate data visualization method and report format are selected according to user needs and preferences.

[0025] Preferably, in the data collection step, data quality monitoring technology is used to monitor the data collection process and data quality in real time, and when data anomalies or quality problems are found, alarms are promptly issued and corresponding processing measures are taken.

[0026] Preferably, in the model training step, model evaluation and optimization techniques are used to regularly evaluate and test the trained model, adjust the model parameters and structure based on the evaluation results, and continuously update and expand the training data.

[0027] Preferably, an electronic device comprises a processor and a memory, wherein the memory stores a computer program, and the processor implements the large model-based clinical trial data analysis method when executing the computer program.

[0028] Preferably, a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the clinical trial data analysis method based on a large model is implemented.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] 1. Efficient data collection and integration:

[0031] Through advanced data interface technology, data can be automatically collected quickly and accurately from various clinical trial data sources, greatly improving the efficiency of data collection. At the same time, it supports automatic conversion and analysis of multiple data formats, solves the data compatibility problem between different data sources, and realizes efficient data integration. For example, it can seamlessly integrate electronic medical record systems, laboratory information management systems, and medical equipment data into a unified data platform, providing a comprehensive and accurate data foundation for subsequent analysis.

[0032] 2. Intelligent data processing and quality improvement:

[0033] The use of intelligent data cleaning algorithms can automatically identify and remove outliers, duplicate data, and erroneous data in the data, thereby improving the quality of the data. The application of data cleaning effect evaluation formulas allows the effect of data cleaning to be quantified and monitored, ensuring the accuracy and reliability of the data. The standardized data format conversion method converts data from different sources into a standard format, facilitating subsequent analysis and processing.

[0034] 3. Powerful model building and analysis capabilities:

[0035] The use of advanced deep learning algorithms and distributed training technology to build large models can effectively capture the complex relationships and patterns in clinical trial data. The application of model training convergence indicators ensures the effectiveness and stability of model training. In terms of efficacy evaluation, safety analysis, and patient characteristics analysis, it can provide accurate and comprehensive analysis results, providing a scientific basis for clinical decision-making. For example, through efficacy evaluation indicators and safety evaluation indicators, the efficacy and safety of drugs can be accurately evaluated to help doctors choose the best treatment plan.

[0036] 4. Flexible and diverse results display and interpretation:

[0037] It supports a variety of data visualization methods, such as charts, reports, and dynamic images, and can present complex analysis results to users in an intuitive and easy-to-understand form. The application of result display satisfaction evaluation indicators allows the result display method to be continuously optimized based on user feedback to improve user satisfaction. At the same time, the automatically generated detailed result explanations and clinical recommendations help users better understand and apply the analysis results, improving the efficiency and quality of decision-making.

[0038] 5. Real-time update and dynamic optimization:

[0039] The use of real-time data analysis and dynamic model update technology can update and optimize the analysis model in real time based on newly collected data and user feedback. This enables the system to always maintain accurate analysis and prediction capabilities for clinical trial data and adapt to the ever-changing clinical trial environment and data characteristics. For example, when new clinical trial data is generated or user needs change, the system can adjust the model and analysis strategy in a timely manner to provide the latest analysis results.

[0040] 6. Personalized service and customization:

[0041] The result output module supports personalized result customization and push services. Users can choose specific analysis results and data display methods according to their needs, and set the time and method of automatic push. This meets the personalized needs of different users and improves the user experience and work efficiency. For example, doctors can customize the analysis results of the patient groups they are concerned about, and the system will regularly push the latest analysis reports to the doctor's mailbox or mobile device, so that the doctor can understand the patient's condition at any time. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a schematic block diagram of a large model-based clinical trial data analysis system proposed by the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0045] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" 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 a direct connection, or it can be indirectly connected through an intermediate medium, and it can be the internal connection of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0046] Reference Figure 1 : A clinical trial data analysis system based on a large model, including a data acquisition module, a data preprocessing module, a model training module, a data analysis module and a result output module;

[0047] The data collection module is used to efficiently collect raw data from various clinical trial data sources, including electronic medical record systems, laboratory information management systems, medical equipment data, etc. It uses advanced data interface technology and secure data encryption transmission methods to ensure data accuracy and completeness; for example, it supports automatic conversion and parsing of multiple data formats. Data collection efficiency evaluation indicators Where E collect represents the data collection efficiency, N collected Indicates the amount of data collected per unit time, T collect Indicates the collection time;

[0048] The data preprocessing module cleans, standardizes, and annotates the collected data. It uses intelligent data cleaning algorithms to automatically identify and remove outliers, duplicate data, and erroneous data. It uses standardized data format conversion methods to unify data formats and accurately annotate data based on professional medical knowledge and annotation rules. The data cleaning effect evaluation formula is: Where Q represents the quality improvement rate after data cleaning, N clean Indicates the amount of data after cleaning, N error Indicates the amount of error data before cleaning, N original Indicates the amount of raw data;

[0049] The model training module uses large-scale clinical trial data to train large models, and uses advanced deep learning algorithms and distributed training technologies to build high-performance data analysis models, such as neural network models based on the Transformer architecture. It optimizes model parameters and structures through large-scale data iterative training. The model training convergence index C = |f(x n )-f(x n-1 )|<∈, where C represents the model convergence condition, f(x n ) and f(x n-1 ) represent the model output values ​​of two adjacent iterations, ∈ is the preset minimum value;

[0050] The data analysis module uses the trained large model to conduct in-depth analysis of the pre-processed data, including efficacy evaluation, safety analysis, and patient characteristics analysis. It mines the potential information of the data through the powerful computing power and intelligent algorithm of the model, providing a scientific basis for clinical trial decision-making and optimization. efficacy =α×P recovery +β×T improve Among them, R efficacy represents the efficacy evaluation value, α and β are weight coefficients, P recovery Represents the patient recovery rate and other related indicators, T inprove Indicates relevant indicators such as the time it takes for patients’ symptoms to improve;

[0051] The result output module displays the analysis results to users in an intuitive and easy-to-understand form, supports a variety of data visualization methods such as charts, reports, dynamic images, etc., and provides detailed data reports and analysis interpretations. Where S represents the user's satisfaction with the result display, N satis fied The number of satisfied users, N total Indicates the total number of users.

[0052] In the present invention, the data interface technology in the data acquisition module supports two modes: real-time data acquisition and batch data acquisition, and can flexibly select the acquisition method according to different application scenarios and data sources. At the same time, data encryption transmission uses the Advanced Encryption Standard AES algorithm to ensure data security.

[0053] In the present invention, the distributed training technology in the model training module adopts a combination of data parallelism and model parallelism. Large-scale clinical trial data is divided into multiple small data blocks and distributed to multiple computing nodes for parallel training. At the same time, different parts of the model are distributed to different nodes for collaborative training to improve training efficiency. The training efficiency improvement index is: Where T represents the percentage of training efficiency improvement, T single Represents the single node training time, T distributed Represents the distributed training time.

[0054] In the present invention, the safety analysis in the data analysis module adopts a multi-dimensional data analysis method. A comprehensive evaluation model is constructed by combining various clinical trial indicators and patient baseline characteristics, for example, analyzing the relationship between factors such as drug dosage, treatment time, patient age, gender, and safety. Safety evaluation index S safety =γ×A adveerse +δ×F frequency , where S safety represents the safety assessment value, γ and δ are weight coefficients, A adverse Indicates the probability of adverse events and other related indicators, F frequency Indicates relevant indicators such as the frequency of adverse events.

[0055] In the present invention, the result output module supports personalized result customization and push service, and users can select specific analysis results and data display methods according to their own needs, and set the time and method of automatic push.

[0056] The present invention also discloses a clinical trial data analysis method based on a large model, comprising the following steps:

[0057] S1. Data collection: Use the data collection module to collect raw data from various clinical trial data sources, and conduct preliminary sorting and classification, and record the data source, time and related clinical trial information;

[0058] S2, data preprocessing: the collected data is transferred to the data preprocessing module for cleaning, standardization, labeling and other processing operations;

[0059] S3. Model training: Input the preprocessed clinical trial data into the model training module, use advanced deep learning algorithms and distributed training technology to train large models, and select appropriate model architecture and parameter settings based on data characteristics and analysis requirements;

[0060] S4. Data analysis: Use the trained large model to conduct in-depth analysis of clinical trial data, and select corresponding analysis algorithms and model parameters based on user needs and set analysis goals;

[0061] S5. Result output: The analysis results are displayed and output through the result output module, and the appropriate data visualization method and report format are selected according to user needs and preferences.

[0062] In the present invention, in the data collection step, data quality monitoring technology is used to monitor the data collection process and data quality in real time. When data anomalies or quality problems are found, an alarm is issued in time and corresponding processing measures are taken.

[0063] In the present invention, in the model training step, model evaluation and optimization techniques are used to regularly evaluate and test the trained model, adjust the model parameters and structure according to the evaluation results, and continuously update and expand the training data.

[0064] The present invention also discloses an electronic device, including a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, the clinical trial data analysis method based on a large model is implemented.

[0065] The present invention also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the clinical trial data analysis method based on a large model is implemented.

[0066] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A clinical trial data analysis system based on a large model, including a data acquisition module, a data preprocessing module, a model training module, a data analysis module and a result output module; Data collection module: used to collect raw data from clinical trial data sources, including electronic medical record systems, laboratory information management systems, and medical equipment data. It uses data interface technology and data encryption transmission methods, supports automatic data format conversion and analysis, and the data collection efficiency evaluation index formula is: Where E collect represents the data collection efficiency, N collected Indicates the amount of data collected per unit time, T collect Indicates the collection time; Data preprocessing module: cleans, standardizes and annotates the collected data, uses data cleaning algorithms to automatically identify and remove outliers, duplicate data and erroneous data, uses standardized data format conversion methods to unify data formats, and annotates data based on medical knowledge and annotation rules. The data cleaning effect evaluation formula is: Where Q represents the quality improvement rate after data cleaning, N clean Indicates the amount of data after cleaning, N error Indicates the amount of error data before cleaning, N original Indicates the amount of raw data; Model training module: Use clinical trial data to train a large model, and use a neural network model based on the Transformer architecture to build a high-performance data analysis model. The model training convergence index C = |f(x n )-f(x n-1 )|<∈, where C represents the model convergence condition, f(x n ) and f(x n-1 ) represent the model output values ​​of two adjacent iterations, ∈ is the preset minimum value; Data analysis module: Use the trained large model to analyze the pre-processed data, including efficacy evaluation, safety analysis, and patient characteristics analysis. The efficacy evaluation index formula is R efficacy =α×P recoυery +β×T improυe , where R efficacy represents the efficacy evaluation value, α and β are weight coefficients, P recoυery Represents the patient recovery rate and other related indicators, T improυe Indicates relevant indicators such as the time it takes for patients’ symptoms to improve; Result output module: Displays the analysis results to users, supports data reports and analysis interpretation in the form of charts, reports or dynamic image data visualization, and displays satisfaction evaluation Where S represents the user's satisfaction with the result display, N satis fied The number of satisfied users, N total Indicates the total number of users.

2. The clinical trial data analysis system based on a large model according to claim 1, characterized in that: The data interface technology in the data acquisition module supports two modes: real-time data acquisition and batch data acquisition. The acquisition method is selected according to different application scenarios and data sources. The data encryption transmission adopts the encryption standard AES algorithm to ensure data security.

3. The clinical trial data analysis system based on a large model according to claim 1, characterized in that: The distributed training technology in the model training module adopts a combination of data parallelism and model parallelism to divide the clinical trial data into small data blocks and distribute them to computing nodes for parallel training. At the same time, different parts of the model are distributed to different nodes for collaborative training. The training efficiency improvement index is Where T represents the percentage of training efficiency improvement, T single Represents the single node training time, T distributed Represents the distributed training time.

4. The clinical trial data analysis system based on a large model according to claim 1, characterized in that: The safety analysis in the data analysis module adopts a multi-dimensional data analysis method, combining various clinical trial indicators and patient baseline characteristics to build a comprehensive evaluation model, including analyzing the relationship between drug dosage, treatment time, patient age and gender factors and safety, and the safety evaluation index S safety =γ×A adυerse +δ×F frequency , where S safety represents the safety assessment value, γ and δ are weight coefficients, A adυerse Indicates the probability of adverse events and other related indicators, F frequency Indicates relevant indicators such as the frequency of adverse events.

5. The clinical trial data analysis system based on a large model according to claim 1, characterized in that: The result output module supports personalized result customization and push services. Users can select specific analysis results and data display methods according to their own needs, and set the time and method of automatic push.

6. A clinical trial data analysis method based on a large model is used to implement the clinical trial data analysis system based on a large model as described in any one of claims 1 to 5, characterized in that: The following steps are involved: S1. Data collection: Use the data collection module to collect raw data from various clinical trial data sources, and conduct preliminary sorting and classification, and record the data source, time and related clinical trial information; S2, data preprocessing: the collected data is transferred to the data preprocessing module for cleaning, standardization and labeling operations; S3. Model training: Input the preprocessed clinical trial data into the model training module, use deep learning algorithms and distributed training technology to train large models, and select appropriate model architecture and parameter settings based on data characteristics and analysis requirements; S4. Data analysis: Use the trained large model to analyze clinical trial data, and select the corresponding analysis algorithm and model parameters according to user needs and set analysis goals; S5. Result output: The analysis results are displayed and output through the result output module, and the appropriate data visualization method and report format are selected according to user needs and preferences.

7. The clinical trial data analysis method based on a large model according to claim 6, characterized in that: In the data collection step, data quality monitoring technology is used to monitor the data collection process and data quality in real time. When data anomalies or quality problems are found, alarms are issued in a timely manner and corresponding processing measures are taken.

8. The clinical trial data analysis method based on a large model according to claim 6, characterized in that: In the model training step, model evaluation and optimization techniques are used to evaluate and test the trained model, adjust the model parameters and structure according to the evaluation results, and continuously update and expand the training data.

9. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the clinical trial data analysis method based on a large model as described in any one of claims 6 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the large model-based clinical trial data analysis method according to any one of claims 6 to 8 is implemented.

Images