Clinical test scheme deviation identification method and management system

By establishing a data storage location label and identification deviation model library for multiple models in clinical trials, detecting and repairing clinical trial data, the problem of deviating deviation recognition accuracy caused by errors in data source and multi-platform interaction is solved, and data analysis is realized Improved accuracy and accuracy of deviation recognition.

CN119028601AActive Publication Date: 2024-11-26GENERAL HOSPITAL OF PLA

Patent Information

Application Number
CN202411110749.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-26
Estimated Expiration
2044-08-14

AI Technical Summary

Technical Problem

In clinical trials, errors occurred during the interaction between the data source and multi-platform data, resulting in the loss of accuracy of data analysis patterns and reducing the accuracy of clinical trial protocol deviation recognition.

Method used

By obtaining clinical trial data, establishing data storage location tags, distinguishing data that can be used and to be repaired, building a database of identification deviation models for multiple models, formulating data detection rules, detecting and repairing data, generating deviation results reports, and ensuring data analysis accuracy and accuracy of deviation recognition.

Benefits of technology

It effectively solves the error problem in data source and multi-platform data interaction, improves the accuracy and transparency of clinical trial protocol deviation recognition, and enhances the efficiency and traceability of the trial.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119028601A_ABST
    Figure CN119028601A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of medical data processing and analysis, and provides a clinical test scheme deviation identification method and management system.The clinical test scheme deviation identification method comprises the steps that a clinical test scheme deviation identification model library is constructed, firstly, data storage position labels are established, use data are stored in a classified mode and simulated, and to-be-repaired and usable data are marked; thirdly, constructing a deviation identification model library and a data detection rule, and generating a data maintenance sequence; and then, processing the data into different to-be-detected data groups, and repairing to obtain the repaired data. And finally, determining a detection sequence through a data detection rule, detecting each group of data by using the model library, comparing results and generating a deviation result. According to the method, the quality of clinical test data is improved, data deviation is effectively recognized and repaired, powerful support is provided for accuracy and reliability of clinical tests, and the safety of patients and the effectiveness of the tests are further guaranteed.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing and analysis, and in particular to a method for identifying deviations from clinical trial protocols and a management system. Background Art

[0002] With the rapid development of science and technology, digital technology has penetrated into all walks of life and brought revolutionary changes to many fields. In the medical field, especially in clinical trials, the application of digital technology is becoming more and more extensive. At present, advanced technologies such as cloud computing, big data, and the Internet of Things provide strong support for data management, monitoring, and analysis of clinical trials, which has significantly improved the efficiency and accuracy of clinical trials.

[0003] In clinical trials, digital means are widely used to identify protocol deviations. Specifically, through big data technology, massive clinical trial data can be analyzed and mined in real time, so as to promptly detect anomalies and deviations in the data. Cloud computing provides powerful computing power and storage space, making data processing and analysis more efficient. In addition, IoT technology can also be used to monitor the physiological data and trial process of subjects in real time to ensure that the trial is carried out strictly in accordance with the protocol. The combined use of these digital means can greatly improve the transparency and traceability of clinical trials, and help to promptly identify and correct protocol deviations.

[0004] However, although digital means have shown great potential in identifying deviations from clinical trial protocols, there are still some technical pain points. First, the diversity and complexity of data sources lead to uneven data quality. Clinical trials involve data interactions across multiple systems and platforms. When errors occur at the source of data and during multi-platform data interactions, the data analysis model trained based on clinical data will lose accuracy, thereby reducing the accuracy of identifying deviations from clinical trial protocols. Summary of the invention

[0005] The purpose of the present invention is to provide a method and management system for identifying deviations from clinical trial protocols, so as to solve the problem that when errors occur at the source of data and in the process of multi-platform data interaction, the data analysis model trained based on clinical data will lose accuracy, thereby reducing the accuracy of identifying deviations from clinical trial protocols.

[0006] In a first aspect, the present invention provides a method for identifying deviations from a clinical trial protocol, comprising:

[0007] Step S101: Acquire clinical trial data, which includes patient information data, clinical trial operation data, clinical observation data, and other external data;

[0008] Step S102: establishing corresponding data storage location labels for each item of data in the clinical trial data, storing each item of data in the clinical trial data in the corresponding data storage location, and performing data simulation use at the respective data storage locations, establishing labels for unusable data to obtain data to be repaired, and establishing labels for usable data to obtain usable data;

[0009] Step S103: construct a clinical trial protocol identification deviation model library, which includes a prediction model, a pattern recognition model, a log analysis model, and a behavior tracking model, establish data detection rules, and generate a data inspection sequence by passing the obtained usable data through the data detection rules;

[0010] Step S104: setting the usable data as the first data group to be detected, setting the data to be repaired and the usable data to form the second data group to be detected, repairing the data to be repaired to obtain repaired data, and setting the repaired data and the usable data to the third data group to be detected;

[0011] Step S105: Establish data detection rules, apply the data detection rules to the first data group to be detected, the second data group to be detected, and the third data group to be detected to obtain a data detection order, retrieve the corresponding detection models in the data detection order in the clinical trial protocol identification deviation model library, detect the first data group to be detected, the second data group to be detected, and the third data to be detected according to the data detection order, and compare the detection results. According to the comparison results, generate the clinical trial protocol deviation results.

[0012] Furthermore, the present invention provides a method for identifying deviations from a clinical trial protocol, step S101: obtaining clinical trial data, the clinical trial data including patient information data, clinical trial operation data, clinical observation data and other external data, including:

[0013] Patient information data: extract data features from medical records to obtain patient medical records, collect patient electronic data through an electronic data collection system, aggregate patient electronic data with patient medical records to obtain patient information data;

[0014] Clinical trial operation data: clinical trial operation data is obtained by converting laboratory records, output information of medical equipment and researchers' records into information data;

[0015] Clinical observation data: obtained through direct observation and recording by doctors or researchers, as well as information automatically recorded by observation equipment;

[0016] Other external data: data information obtained through third-party research institutions and public databases.

[0017] Furthermore, the present invention provides a method for identifying deviations from a clinical trial protocol, step S102: establishing corresponding data storage location labels for various data in the clinical trial data, storing various data in the clinical trial data in corresponding data storage locations, and performing data simulation use at respective data storage locations, establishing labels for unusable data to obtain data to be repaired, and establishing labels for usable data to obtain usable data, including:

[0018] According to the data types in clinical trial data, a data simulation verification knowledge base is established. The data simulation verification knowledge base includes clinical trial simulation usage models, patient information data verification models, clinical observation data simulation usage models, and other external data verification models;

[0019] Establish corresponding data storage location labels for patient information data, clinical trial operation data, clinical observation data and other external data in the clinical trial data. The data storage location labels include folder names and database table names. Each clinical trial data is stored in a corresponding label location to obtain a data container corresponding to the data storage location label.

[0020] Establish a spare data container corresponding to the data container, and perform data verification on the data in the data container through a data simulation verification knowledge base. The data verification includes reading data, performing simple statistical analysis or data query operations;

[0021] When performing a data repair operation, if data is lost or damaged, a spare data container corresponding to the data container is retrieved to replace the lost or damaged data;

[0022] Record problems encountered during simulation use, including read errors, formatting issues, or data inconsistencies;

[0023] According to the results of data verification, the data that cannot be used normally or has problems will be marked as data to be repaired. The data to be repaired includes data format errors, missing key information, and data damage;

[0024] Data that successfully passes data validation and has no problems is marked as usable data.

[0025] Furthermore, the method for identifying deviations from a clinical trial protocol provided by the present invention further includes:

[0026] Creating a data record table to be repaired, the data record table to be repaired includes the file of the data to be repaired, the record data to be repaired and the location information of the data to be repaired;

[0027] Creating an available data record table, which includes files of available data, available record data, and available data location information;

[0028] Dedicated data storage location labels are set for patient information data, clinical trial operation data, clinical observation data and other external data. The data storage location labels include folder names, database table names and corresponding data container identifiers.

[0029] Furthermore, the method for identifying deviations from clinical trial protocols provided by the present invention includes step S103: constructing a model library for identifying deviations from clinical trial protocols, wherein the model library for identifying deviations from clinical trial protocols includes a prediction model, a pattern recognition model, a log analysis model, and a behavior tracking model, establishing data detection rules, and generating a data inspection sequence by passing the obtained usable data through the data detection rules, including:

[0030] Obtain the purpose, design principles, operation procedures and core data points of the clinical trial plan, identify key data indicators for evaluating deviations, and establish data quality standards and scope;

[0031] According to the quality standards and scopes that the data should meet, the data is matched in the historical data to obtain the quality standards and scopes of the data and the corresponding historical data, and the quality standards and scopes of the data and the corresponding historical data are extracted to obtain the data deviation patterns and features;

[0032] Determine the key data indicators for deviation detection, convert the determined detection logic into executable code on the server side, including value range verification, time series comparison, and set up automated scripts to perform deviation detection on newly entered data at preset time periods or in real time;

[0033] The initially established detection rules are verified using a data set with known results, and the rule parameters are adjusted based on the verification feedback. The verified detection rules are integrated into the server-side clinical trial protocol identification deviation model library. During the clinical trial, data changes are monitored in real time through the server-side, and the detection rules are adjusted based on actual conditions.

[0034] Furthermore, the method for identifying deviations from a clinical trial protocol provided by the present invention, step S104: setting the usable data as a first data group to be detected, setting the data to be repaired and the usable data to form a second data group to be detected, repairing the data to be repaired to obtain repaired data, and setting the repaired data and the usable data to a third data group to be detected, comprises:

[0035] The first data group to be detected is established, and the server side selects the records marked as usable data from the database according to the data tags;

[0036] Integrate the available data records into a data set, name the data set as a first data set to be detected, and store the first data set to be detected in a designated server location;

[0037] The second data group to be detected is established, and the server selects the records marked as data to be repaired and usable data according to the data tags;

[0038] The selected data to be repaired is mixed with the usable data, and the mixed data set contains both normal data and data to be repaired;

[0039] Setting the mixed data set as a second data group to be detected, and storing the second data group to be detected in a designated server location;

[0040] The third is to establish a data group to be detected, and the server side calls a data repair algorithm to perform repair processing on the data to be repaired, and the repair processing includes filling missing values ​​and correcting erroneous data;

[0041] After the repair is completed, the repaired data is verified. After the verification is passed, the repaired data is integrated with the original usable data to form a new data set. This new integrated data set is set as the third data set to be tested, and the third data set to be tested is stored in a specified location on the server.

[0042] Furthermore, the present invention provides a method for identifying deviations from a clinical trial protocol, step S105: establishing a data detection rule, applying the data detection rule to the first data group to be detected, the second data group to be detected, and the third data group to be detected to obtain a data detection sequence, retrieving each detection model corresponding to the data detection sequence in the clinical trial protocol identification deviation model library, detecting the first data group to be detected, the second data group to be detected, and the third data group to be detected according to the data detection sequence, and comparing the detection results, and generating a clinical trial protocol deviation result according to the comparison result, including:

[0043] According to the established data detection rules, the first to-be-detected data group, the second to-be-detected data group and the third to-be-detected data group are detected one by one;

[0044] For the first data group to be tested, the second data group to be tested, and the third data group to be tested, corresponding detection models are retrieved from the clinical trial protocol identification deviation model library, and detection is performed according to a predetermined data detection order;

[0045] After the detection is completed, the detection results corresponding to the first data group to be detected, the second data group to be detected, and the third data group to be detected are collected and recorded. The detection results include the identified potential deviation points, detailed information of the abnormal data, and the severity of the deviation. By comparison, common deviation points and deviation points in specific data groups are identified;

[0046] Analyze the improvement of deviation identification on the repaired data in the third data group to be tested and the original data in the first data group to be tested and the second data group to be tested, and generate a clinical trial protocol deviation result report based on the comparison results,

[0047] The clinical trial protocol deviation result report includes the deviation points identified in each group of data, including the type of deviation, the cause of the deviation, and the degree of impact of the deviation, as well as the impact of the repair on the accuracy of deviation identification. The generated deviation result report is output to the specified location.

[0048] Furthermore, the method for identifying deviations from the clinical trial protocol provided by the present invention analyzes the repaired data in the third data group to be tested and the improvement in deviation identification of the original data in the first data group to be tested and the second data group to be tested, including:

[0049] Compare the number of deviation points in the first data group to be detected, the second data group to be detected and the third data group to be detected. If the number of deviation points of the repaired data in the third data group to be detected is reduced, the repair operation is effective;

[0050] Comparing the repaired data in the third data group to be detected with the original data in the first data group to be detected and the second data group to be detected;

[0051] By comparing the data before and after the repair, if the number of deviating points in the repaired data is reduced, the repair effect is positive;

[0052] By comparing the recognition accuracy of deviation points in the data before and after repair, the improvement of deviation recognition accuracy caused by the repair operation can be evaluated. If the repaired data has a high accuracy in deviation recognition, the repair operation reduces the number of deviation points and improves the accuracy of deviation recognition in the clinical trial protocol.

[0053] In a second aspect, a clinical trial protocol deviation identification and management system uses the clinical trial protocol deviation identification method, including a server side, a background control side and a user side, the server side establishes a communication connection with the background control side and the user side, and the server side includes:

[0054] The server side obtains clinical trial data, which includes patient information data, clinical trial operation data, clinical observation data and other external data;

[0055] The server creates corresponding data storage location labels for each data in the clinical trial data, stores each data in the clinical trial data in the corresponding data storage location, and simulates the use of data in each data storage location, creates labels for unusable data to obtain data to be repaired, and creates labels for usable data to obtain usable data;

[0056] The server side builds a clinical trial protocol identification deviation model library, which includes prediction models, pattern recognition models, log analysis models, and behavior tracking models. It also builds data detection rules, and generates a data inspection sequence by passing the available data through the data detection rules.

[0057] The server sets the usable data as the first data group to be detected, sets the data to be repaired and the usable data to form the second data group to be detected, repairs the data to be repaired to obtain repaired data, and sets the repaired data and the usable data to the third data group to be detected;

[0058] The server side establishes data detection rules, applies the first data group to be detected, the second data group to be detected, and the third data group to be detected to the data detection rules, obtains the data detection sequence, retrieves the corresponding detection models in the data detection sequence in the clinical trial protocol identification deviation model library, detects the first data group to be detected, the second data group to be detected, and the third data to be detected according to the data detection sequence, and compares the detection results. Based on the comparison results, the clinical trial protocol deviation result is generated, and the server side sends the clinical trial protocol deviation result to the user side.

[0059] The beneficial effects of the present invention are as follows: The present invention provides a method and management system for identifying deviations from clinical trial protocols. The present invention combines multiple models to effectively solve errors in data sources and multi-platform data interactions, ensure the accuracy of data analysis, and improve the accuracy of deviation identification. At the same time, by analyzing clinical trial data in real time and monitoring subject information, the transparency and traceability of clinical trials are enhanced, which helps to identify and correct deviations in a timely manner, thereby improving trial efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0061] Figure 1 A schematic flow chart of a method for identifying deviations from a clinical trial protocol provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described in conjunction with the specific embodiments of the present invention and the corresponding drawings. 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. The technical solutions provided by the embodiments of the present invention are described in detail below in conjunction with the drawings.

[0063] First, see Figure 1 , the present invention provides a method for identifying deviations from a clinical trial protocol, comprising:

[0064] In step S101: clinical trial data is obtained, the clinical trial data includes patient information data, clinical trial operation data, clinical observation data and other external data;

[0065] The server needs to clearly identify the source of clinical trial data. Data comes from multiple channels, including but not limited to:

[0066] Patient information data: mainly comes from the hospital's medical record system, electronic health record (EHR) or information directly input by the patient. Specifically includes key information such as the patient's age, gender, medical history, family medical history, etc.

[0067] Clinical trial operational data are usually automatically generated and recorded by laboratory technicians, researchers or medical equipment, including the dosage of trial drugs, administration time, laboratory test results, etc.

[0068] Clinical observation data are directly observed and recorded by doctors and researchers based on the patients' clinical manifestations and reactions, such as symptom changes, physical signs and reactions, etc.

[0069] Other external data comes from third-party research institutions, public databases or relevant literature to assist in analysis and comparison.

[0070] The server needs to establish a stable and secure data interface to obtain data from the above sources, specifically involving the following aspects:

[0071] Data format standardization ensures that data obtained from different sources can be stored and processed in a unified format.

[0072] Data encryption and secure transmission: Use encryption technologies such as SSL / TLS to ensure the security of data transmission.

[0073] Data verification and cleaning: Before the data is received, preliminary data verification is performed to eliminate invalid or erroneous data to ensure the accuracy and completeness of the data.

[0074] Data storage preparation. Before data is officially stored, the server needs to do the following preparations: Design a reasonable database model based on data type and structure, including data tables, fields, indexes, etc. Estimate the amount of data, allocate storage space reasonably, and ensure data scalability and accessibility. Develop a data backup plan to prevent data loss or damage, and ensure that data can be quickly restored when necessary.

[0075] In the data acquisition process, the server sends requests to various data sources to transmit relevant data. After receiving the data, the server performs preliminary verification and cleaning to ensure the accuracy and validity of the data. The cleaned data is stored in a pre-designed database for subsequent processing and analysis.

[0076] Logging and monitoring: During the entire data acquisition process, the server also needs to log and monitor performance, recording the specific time and content of each operation so that problems can be quickly located and investigated. Real-time monitoring of data transmission speed, stability, and integrity ensures the smooth progress of the data acquisition process.

[0077] In step S102: a corresponding data storage location label is established for each data in the clinical trial data, each data in the clinical trial data is stored in the corresponding data storage location, and data simulation is performed at each data storage location, a label is established for unusable data to obtain data to be repaired, and a label is established for usable data to obtain usable data;

[0078] The server first carefully classifies the acquired clinical trial data, including but not limited to basic patient information, trial operation records, observation data, etc. Each type of data should be clearly defined and identified.

[0079] For each type of data, the server will assign a specific data storage location label to it. The label not only indicates the storage path of the data, but also includes the type of data and how to use it.

[0080] Using the previously established data storage location labels, various types of data are accurately written to the corresponding storage locations. In this process, the integrity and consistency of the data must be ensured.

[0081] In order to verify the availability of data, the server needs to build a simulated usage environment, which should be as close to the actual usage scenario as possible. In the simulated environment, the server will try to read and use the data stored in various locations. The purpose of this step is to detect whether the data has format errors, missing values, outliers, and other problems. Based on the results of the simulated use, the server will determine which data is available and which data has problems. For data that cannot be used, the server will label it as "to be repaired"; for data that can be used normally, it will be labeled "available for use".

[0082] For data marked as "to be repaired", the server will record its detailed information and problem for subsequent repair operations. At the same time, this data will be temporarily isolated to prevent it from interfering with subsequent analysis. For data marked as "usable", the server will further organize and optimize it to improve its efficiency in subsequent use. This data will be included in the subsequent data analysis process.

[0083] During the entire process, the server will record the operation and results of each step in detail so that problems can be quickly located and resolved. For abnormal situations that occur during the simulation, such as data reading failure, format errors, etc., the server will have a complete exception handling mechanism to deal with them to ensure the smooth progress of the entire process.

[0084] In step S103: construct a clinical trial protocol identification deviation model library, which includes a prediction model, a pattern recognition model, a log analysis model, and a behavior tracking model, establish data detection rules, and use the obtained usable data through the data detection rules to generate a data inspection sequence;

[0085] According to the specific needs of clinical trials and historical data analysis, select appropriate models to build a deviation identification model library. These models include but are not limited to prediction models, pattern recognition models, log analysis models, and behavior tracking models.

[0086] Use historical data to train predictive models to forecast specific metrics or outcomes and identify deviations from expected results.

[0087] Pattern recognition models: Pattern recognition models are trained through machine learning algorithms to discover hidden patterns and anomalies in data.

[0088] Log analysis model: Analyze system logs and user behavior logs to identify abnormal operations or system errors.

[0089] Behavior tracking model: Tracks the behavior sequence of users or systems and detects abnormal behavior patterns.

[0090] Model integration and testing: Integrate the trained models into the deviation identification model library and conduct comprehensive testing to ensure the collaboration and accuracy of the models.

[0091] Develop a comprehensive set of data detection rules based on clinical trial norms, historical deviation cases, and expert knowledge. These rules are designed to identify anomalies, errors, or patterns that do not conform to expectations in the data. Test these rules on actual data, and adjust and optimize the rules based on the test results to improve the accuracy and efficiency of deviation identification.

[0092] Perform necessary preprocessing on the available data, such as standardization and denoising, to ensure data quality and meet the model input requirements. Data passes detection rules: The preprocessed data passes the previously established data detection rules to identify deviation points. Based on the output results of the detection rules, a data repair sequence is generated. This sequence indicates the data points and related models that need further inspection, repair or verification.

[0093] In step S104: the usable data is set as the first data group to be detected, the data to be repaired and the usable data are set to form a second data group to be detected, the data to be repaired is repaired to obtain repaired data, and the repaired data and the usable data are set as the third data group to be detected;

[0094] First data group to be tested: The server first sets the data previously marked as "usable" as the first data group to be tested. These data have been confirmed to be accurate and reliable in the previous verification process, so they can be directly used for subsequent testing and analysis.

[0095] Second data group to be tested: Subsequently, the server mixes the data to be repaired with the usable data to form the second data group to be tested. The purpose of this group is to observe the impact of the data to be repaired on the overall data quality in subsequent tests and try to identify potential problems of these data in the mixed environment.

[0096] Data repair strategy: The server checks the data to be repaired one by one according to the previously established data detection rules and the problems identified in the simulation. For different types of errors or anomalies (such as format errors, missing values, logical errors, etc.), the server will adopt corresponding repair strategies, such as data interpolation, format conversion, logical verification, etc.

[0097] Data repair execution: During the repair process, the server will record each step of the repair operation and results to ensure data traceability and repair accuracy. After the repair is completed, the data will be considered "repaired data".

[0098] Finally, the server merges the repaired data with the original usable data and sets it as the third set of data to be tested. This set of data will be used for subsequent comprehensive testing and verification to ensure the effectiveness of the repair operation and the overall quality of the data.

[0099] During the entire data grouping and repair process, the server will record the operations and results of each step in detail, including the specific methods of repair, data comparison before and after repair, etc. This log information is crucial for subsequent data audits and problem tracking. The server will regularly conduct quality control checks on the grouping and repair process to ensure that each step meets established standards and specifications. This helps to identify problems in a timely manner and make adjustments, thereby continuously improving the accuracy and efficiency of data processing.

[0100] In step S105: establish data detection rules, apply the data detection rules to the first data group to be detected, the second data group to be detected and the third data group to be detected, obtain the data detection sequence, retrieve the corresponding detection models in the data detection sequence in the clinical trial protocol identification deviation model library, detect the first data group to be detected, the second data group to be detected and the third data to be detected according to the data detection sequence, compare the detection results, and generate the clinical trial protocol deviation results based on the comparison results.

[0101] Based on the specific needs and historical data of clinical trials, further refine and improve the data detection rules. Before actual application, verify the formulated data detection rules to ensure their effectiveness and applicability. This can be done by testing on a small-scale data set.

[0102] Data grouping passes detection rules: the first data group to be detected, the second data group to be detected and the third data group to be detected pass the data detection rules respectively.

[0103] The order of testing each group of data is determined based on the importance of the data and the urgency of repair. For example, the repaired data (the third group of data to be tested) is tested first to verify the repair effect, followed by the original usable data (the first group of data to be tested), and finally the data group containing the data to be repaired (the second group of data to be tested).

[0104] According to the data detection sequence, the corresponding detection model is retrieved from the clinical trial protocol identification deviation model library. These models include prediction models, pattern recognition models, log analysis models, and behavior tracking models.

[0105] In accordance with the determined test order, use the retrieved model to test each set of data. During the test process, the integrity and consistency of the data should be ensured, and any potential deviations or anomalies should be discovered and recorded in a timely manner.

[0106] The test results of each group of data were compared and analyzed, with special attention paid to the differences between the repaired data and the original data and the impact of these differences on the clinical trial plan.

[0107] Based on the comparison results, the clinical trial protocol deviation results are generated. The clinical trial protocol deviation results include information such as the specific deviation points identified, the type and degree of deviation, and the impact on the clinical trial protocol.

[0108] Specifically, the method for identifying deviations from a clinical trial protocol provided by the present invention includes step S101: obtaining clinical trial data, where the clinical trial data includes patient information data, clinical trial operation data, clinical observation data, and other external data, including:

[0109] Patient information data: extract data features from medical records to obtain patient medical records, collect patient electronic data through an electronic data collection system, aggregate patient electronic data with patient medical records to obtain patient information data;

[0110] Clinical trial operation data: clinical trial operation data is obtained by converting laboratory records, output information of medical equipment and researchers' records into information data;

[0111] Clinical observation data: obtained through direct observation and recording by doctors or researchers, as well as information automatically recorded by observation equipment;

[0112] Other external data: data information obtained through third-party research institutions and public databases.

[0113] Specifically, the method for identifying deviations from a clinical trial protocol provided by the present invention, step S102: establishing corresponding data storage location labels for various data in the clinical trial data, storing various data in the clinical trial data in corresponding data storage locations, and performing data simulation use at respective data storage locations, establishing labels for unusable data to obtain data to be repaired, and establishing labels for usable data to obtain usable data, including:

[0114] According to the data types in clinical trial data, a data simulation verification knowledge base is established. The data simulation verification knowledge base includes clinical trial simulation usage models, patient information data verification models, clinical observation data simulation usage models, and other external data verification models;

[0115] Establish corresponding data storage location labels for patient information data, clinical trial operation data, clinical observation data and other external data in the clinical trial data. The data storage location labels include folder names and database table names. Each clinical trial data is stored in a corresponding label location to obtain a data container corresponding to the data storage location label.

[0116] Establish a spare data container corresponding to the data container, and perform data verification on the data in the data container through a data simulation verification knowledge base. The data verification includes reading data, performing simple statistical analysis or data query operations;

[0117] When performing a data repair operation, if data is lost or damaged, a spare data container corresponding to the data container is retrieved to replace the lost or damaged data;

[0118] Record problems encountered during simulation use, including read errors, formatting issues, or data inconsistencies;

[0119] According to the results of data verification, the data that cannot be used normally or has problems will be marked as data to be repaired. The data to be repaired includes data format errors, missing key information, and data damage;

[0120] Data that successfully passes data validation and has no problems is marked as usable data.

[0121] Specifically, the method for identifying deviations from a clinical trial protocol provided by the present invention further includes:

[0122] Creating a data record table to be repaired, the data record table to be repaired includes the file of the data to be repaired, the record data to be repaired and the location information of the data to be repaired;

[0123] Creating an available data record table, which includes files of available data, available record data, and available data location information;

[0124] Dedicated data storage location labels are set for patient information data, clinical trial operation data, clinical observation data and other external data. The data storage location labels include folder names, database table names and corresponding data container identifiers.

[0125] Specifically, the method for identifying deviations from a clinical trial protocol provided by the present invention includes step S103: constructing a model library for identifying deviations from a clinical trial protocol, wherein the model library for identifying deviations from a clinical trial protocol includes a prediction model, a pattern recognition model, a log analysis model, and a behavior tracking model, establishing data detection rules, and generating a data inspection sequence by passing the obtained usable data through the data detection rules, including:

[0126] Obtain the purpose, design principles, operation procedures and core data points of the clinical trial plan, identify key data indicators for evaluating deviations, and establish data quality standards and scope;

[0127] According to the quality standards and scopes that the data should meet, the data is matched in the historical data to obtain the quality standards and scopes of the data and the corresponding historical data, and the quality standards and scopes of the data and the corresponding historical data are extracted to obtain the data deviation patterns and features;

[0128] Determine the key data indicators for deviation detection, convert the determined detection logic into executable code on the server side, including value range verification, time series comparison, and set up automated scripts to perform deviation detection on newly entered data at preset time periods or in real time;

[0129] The initially established detection rules are verified using a data set with known results, and the rule parameters are adjusted based on the verification feedback. The verified detection rules are integrated into the server-side clinical trial protocol identification deviation model library. During the clinical trial, data changes are monitored in real time through the server-side, and the detection rules are adjusted based on actual conditions.

[0130] Specifically, the method for identifying deviations from a clinical trial protocol provided by the present invention, step S104: setting the usable data as a first data group to be detected, setting the data to be repaired and the usable data to form a second data group to be detected, repairing the data to be repaired to obtain repaired data, and setting the repaired data and the usable data to a third data group to be detected, includes:

[0131] The first data group to be detected is established, and the server side selects the records marked as usable data from the database according to the data tags;

[0132] Integrate the available data records into a data set, name the data set as a first data set to be detected, and store the first data set to be detected in a designated server location;

[0133] The second data group to be detected is established, and the server selects the records marked as data to be repaired and usable data according to the data tags;

[0134] The selected data to be repaired is mixed with the usable data, and the mixed data set contains both normal data and data to be repaired;

[0135] Setting the mixed data set as a second data group to be detected, and storing the second data group to be detected in a designated server location;

[0136] The third is to establish a data group to be detected, and the server side calls a data repair algorithm to perform repair processing on the data to be repaired, and the repair processing includes filling missing values ​​and correcting erroneous data;

[0137] After the repair is completed, the repaired data is verified. After the verification is passed, the repaired data is integrated with the original usable data to form a new data set. This new integrated data set is set as the third data set to be tested, and the third data set to be tested is stored in a specified location on the server.

[0138] Specifically, the method for identifying deviations from a clinical trial protocol provided by the present invention includes step S105: establishing a data detection rule, applying the data detection rule to the first data group to be detected, the second data group to be detected, and the third data group to be detected to obtain a data detection sequence, retrieving each detection model corresponding to the data detection sequence in the clinical trial protocol identification deviation model library, detecting the first data group to be detected, the second data group to be detected, and the third data group to be detected according to the data detection sequence, and comparing the detection results. According to the comparison result, generating a clinical trial protocol deviation result, including:

[0139] According to the established data detection rules, the first to-be-detected data group, the second to-be-detected data group and the third to-be-detected data group are detected one by one;

[0140] For the first data group to be tested, the second data group to be tested, and the third data group to be tested, corresponding detection models are retrieved from the clinical trial protocol identification deviation model library, and detection is performed according to a predetermined data detection order;

[0141] After the detection is completed, the detection results corresponding to the first data group to be detected, the second data group to be detected, and the third data group to be detected are collected and recorded. The detection results include the identified potential deviation points, detailed information of the abnormal data, and the severity of the deviation. By comparison, common deviation points and deviation points in specific data groups are identified;

[0142] Analyze the improvement of deviation identification on the repaired data in the third data group to be tested and the original data in the first data group to be tested and the second data group to be tested, and generate a clinical trial protocol deviation result report based on the comparison results,

[0143] The clinical trial protocol deviation result report includes the deviation points identified in each group of data, including the type of deviation, the cause of the deviation, and the degree of impact of the deviation, as well as the impact of the repair on the accuracy of deviation identification. The generated deviation result report is output to the specified location.

[0144] Specifically, the method for identifying deviations from a clinical trial protocol provided by the present invention analyzes the repaired data in the third data group to be tested and the improvement in deviation identification of the original data in the first data group to be tested and the second data group to be tested, including:

[0145] Compare the number of deviation points in the first data group to be detected, the second data group to be detected and the third data group to be detected. If the number of deviation points of the repaired data in the third data group to be detected is reduced, the repair operation is effective;

[0146] Comparing the repaired data in the third data group to be detected with the original data in the first data group to be detected and the second data group to be detected;

[0147] By comparing the data before and after the repair, if the number of deviating points in the repaired data is reduced, the repair effect is positive;

[0148] By comparing the recognition accuracy of deviation points in the data before and after repair, the improvement of deviation recognition accuracy caused by the repair operation can be evaluated. If the repaired data has a high accuracy in deviation recognition, the repair operation reduces the number of deviation points and improves the accuracy of deviation recognition in the clinical trial protocol.

[0149] In a second aspect, a clinical trial protocol deviation identification and management system uses the clinical trial protocol deviation identification method, including a server side, a background control side and a user side, the server side establishes a communication connection with the background control side and the user side, and the server side includes:

[0150] The server side obtains clinical trial data, which includes patient information data, clinical trial operation data, clinical observation data and other external data;

[0151] The server creates corresponding data storage location labels for each data in the clinical trial data, stores each data in the clinical trial data in the corresponding data storage location, and simulates the use of data in each data storage location, creates labels for unusable data to obtain data to be repaired, and creates labels for usable data to obtain usable data;

[0152] The server side builds a clinical trial protocol identification deviation model library, which includes prediction models, pattern recognition models, log analysis models, and behavior tracking models. It also builds data detection rules, and generates a data inspection sequence by passing the available data through the data detection rules.

[0153] The server sets the usable data as the first data group to be detected, sets the data to be repaired and the usable data to form the second data group to be detected, repairs the data to be repaired to obtain repaired data, and sets the repaired data and the usable data to the third data group to be detected;

[0154] The server side establishes data detection rules, applies the first data group to be detected, the second data group to be detected, and the third data group to be detected to the data detection rules, obtains the data detection sequence, retrieves the corresponding detection models in the data detection sequence in the clinical trial protocol identification deviation model library, detects the first data group to be detected, the second data group to be detected, and the third data to be detected according to the data detection sequence, and compares the detection results. Based on the comparison results, the clinical trial protocol deviation result is generated, and the server side sends the clinical trial protocol deviation result to the user side.

[0155] The present invention effectively solves the error problem that occurs during the interaction between the data source and the multi-platform data. The error problem that occurs during the interaction between the data source and the multi-platform data will cause accuracy loss to the data analysis model based on clinical data training, thereby affecting the accuracy of clinical trial protocol deviation identification. First, the present invention ensures the accurate storage and consistency of data by establishing data storage location labels. Then, through data simulation use, the usable and to-be-repaired data are distinguished, and the problem data is processed. Then, a recognition deviation model library containing multiple models is constructed to conduct in-depth detection of clinical data. In addition, data detection rules are formulated to compare multiple groups of data to verify the accuracy of the data. Finally, the effectiveness of the repair operation is confirmed through the repair effect evaluation, thereby improving the accuracy of clinical trial protocol deviation identification. The present invention ensures the accuracy of the data analysis model and the reliability of clinical trial protocol deviation identification.

[0156] The specific implementation and use of the present invention in the digital medical diagnosis scenario is as follows:

[0157] Data acquisition and preprocessing,First, the system will collect clinical trial data from various sources, including patient information, clinical trial operation records, clinical observation data, etc. Then, a specific data storage location label is assigned to each data to ensure structured storage of the data. In this process, the quality of the data is also verified, and unusable data is marked as to be repaired, and usable data is marked as usable.

[0158] Model library construction and data detection: The system will build a clinical trial protocol deviation identification model library, which contains a variety of models such as prediction models and pattern recognition models, for in-depth analysis and detection of data. At the same time, a set of data detection rules will be formulated to conduct preliminary detection on the data previously marked as usable and generate a data inspection sequence.

[0159] Data grouping and repair: In the data grouping stage, the system sets the usable data as the first data group to be detected, and mixes the data to be repaired with the usable data to form the second data group to be detected. Then, the data to be repaired is repaired, and the repaired data and the usable data together form the third data group to be detected.

[0160] Data detection and result generation: According to the previously established data detection rules, the three sets of data are tested in detail and the test results are compared. The system will generate a clinical trial protocol deviation result report based on the comparison results. The report will list in detail the deviation points identified in each set of data, as well as their types, causes and impacts.

[0161] Result output and application: This deviation result report will be output to a designated location for reference by relevant personnel such as researchers, doctors or regulatory agencies. In digital medical diagnosis scenarios, this report can be used to monitor the progress of clinical trials in real time, detect and correct deviations in a timely manner, and ensure the accuracy and reliability of trial data.

[0162] System architecture,The present invention also proposes a complete clinical trial protocol deviation identification management system architecture,,including a server side, a background control side and a user side.,The server side is responsible for core tasks such as data processing,,model building and result generation, while the user side provides researchers with,intuitive result display and reports.

[0163] Through the above steps, the present invention can achieve effective identification and management of clinical trial data deviations in digital medical diagnosis scenarios, thereby significantly improving the accuracy and reliability of the tests.

[0164] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution in the embodiments of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention or some parts of the embodiments.

[0165] The above-described embodiments of the present invention do not limit the protection scope of the present invention.

Claims

1. A method for identifying deviations from a clinical trial protocol, characterized in that: include; Step S101: Acquire clinical trial data, which includes patient information data, clinical trial operation data, clinical observation data, and other external data; Step S102: establishing corresponding data storage location labels for each item of data in the clinical trial data, storing each item of data in the clinical trial data in the corresponding data storage location, and performing data simulation use at the respective data storage locations, establishing labels for unusable data to obtain data to be repaired, and establishing labels for usable data to obtain usable data; Step S103: construct a clinical trial protocol identification deviation model library, which includes a prediction model, a pattern recognition model, a log analysis model, and a behavior tracking model, establish data detection rules, and generate a data inspection sequence by passing the obtained usable data through the data detection rules; Step S104: setting the usable data as the first data group to be detected, setting the data to be repaired and the usable data to form the second data group to be detected, repairing the data to be repaired to obtain repaired data, and setting the repaired data and the usable data to the third data group to be detected; Step S105: Establish data detection rules, apply the data detection rules to the first data group to be detected, the second data group to be detected, and the third data group to be detected to obtain a data detection order, retrieve the corresponding detection models in the data detection order in the clinical trial protocol identification deviation model library, detect the first data group to be detected, the second data group to be detected, and the third data to be detected according to the data detection order, and compare the detection results. According to the comparison results, generate the clinical trial protocol deviation results.

2. The method for identifying deviation from a clinical trial protocol according to claim 1, characterized in that: Step S101: Acquire clinical trial data, which includes patient information data, clinical trial operation data, clinical observation data, and other external data, including: Patient information data: extract data features from medical records to obtain patient medical records, collect patient electronic data through an electronic data collection system, aggregate patient electronic data with patient medical records to obtain patient information data; Clinical trial operation data: clinical trial operation data is obtained by converting laboratory records, output information of medical equipment and researchers' records into information data; Clinical observation data: obtained through direct observation and recording by doctors or researchers, as well as information automatically recorded by observation equipment; Other external data: data information obtained through third-party research institutions and public databases.

3. The method for identifying deviation from a clinical trial protocol according to claim 1, characterized in that: Step S102: Establish corresponding data storage location labels for each data in the clinical trial data, store each data in the clinical trial data in the corresponding data storage location, and simulate the use of data in each data storage location, establish labels for unusable data to obtain data to be repaired, and establish labels for usable data to obtain usable data, including: According to the data types in clinical trial data, a data simulation verification knowledge base is established. The data simulation verification knowledge base includes clinical trial simulation usage models, patient information data verification models, clinical observation data simulation usage models, and other external data verification models; Establish corresponding data storage location labels for patient information data, clinical trial operation data, clinical observation data and other external data in the clinical trial data. The data storage location labels include folder names and database table names. Each clinical trial data is stored in a corresponding label location to obtain a data container corresponding to the data storage location label. Establish a spare data container corresponding to the data container, and perform data verification on the data in the data container through a data simulation verification knowledge base. The data verification includes reading data, performing simple statistical analysis or data query operations; When performing a data repair operation, if data is lost or damaged, a spare data container corresponding to the data container is retrieved to replace the lost or damaged data; Record problems encountered during simulation use, including read errors, formatting issues, or data inconsistencies; According to the results of data verification, the data that cannot be used normally or has problems will be marked as data to be repaired. The data to be repaired includes data format errors, missing key information, and data damage; Data that successfully passes data validation and has no problems is marked as usable data.

4. The method for identifying deviations from a clinical trial protocol according to claim 1, wherein: Also includes: Creating a data record table to be repaired, the data record table to be repaired includes the file of the data to be repaired, the record data to be repaired and the location information of the data to be repaired; Creating an available data record table, which includes files of available data, available record data, and available data location information; Dedicated data storage location labels are set for patient information data, clinical trial operation data, clinical observation data and other external data. The data storage location labels include folder names, database table names and corresponding data container identifiers.

5. The method for identifying deviation from a clinical trial protocol according to claim 1, wherein: Step S103: construct a clinical trial protocol identification deviation model library, which includes a prediction model, a pattern recognition model, a log analysis model, and a behavior tracking model. Establish data detection rules, and use the obtained usable data through the data detection rules to generate a data inspection sequence, including: Obtain the purpose, design principles, operation procedures and core data points of the clinical trial plan, identify key data indicators for evaluating deviations, and establish data quality standards and scope; According to the quality standards and scopes that the data should meet, the data is matched in the historical data to obtain the quality standards and scopes of the data and the corresponding historical data, and the quality standards and scopes of the data and the corresponding historical data are extracted to obtain the data deviation patterns and features; Determine the key data indicators for deviation detection, convert the determined detection logic into executable code on the server side, including value range verification, time series comparison, and set up automated scripts to perform deviation detection on newly entered data at preset time periods or in real time; The initially established detection rules are verified using a data set with known results, and the rule parameters are adjusted based on the verification feedback. The verified detection rules are integrated into the server-side clinical trial protocol identification deviation model library. During the clinical trial, data changes are monitored in real time through the server-side, and the detection rules are adjusted based on actual conditions.

6. The method for identifying deviation from a clinical trial protocol according to claim 1, wherein: Step S104: setting the usable data as the first data group to be detected, setting the data to be repaired and the usable data to form a second data group to be detected, repairing the data to be repaired to obtain repaired data, and setting the repaired data and the usable data to form a third data group to be detected, including: The first data group to be detected is established, and the server side selects the records marked as usable data from the database according to the data tags; Integrate the available data records into a data set, name the data set as a first data set to be detected, and store the first data set to be detected in a designated server location; The second data group to be detected is established, and the server selects the records marked as data to be repaired and usable data according to the data tags; The selected data to be repaired is mixed with the usable data, and the mixed data set contains both normal data and data to be repaired; Setting the mixed data set as a second data group to be detected, and storing the second data group to be detected in a designated server location; The third is to establish a data group to be detected, and the server side calls a data repair algorithm to perform repair processing on the data to be repaired, and the repair processing includes filling missing values ​​and correcting erroneous data; After the repair is completed, the repaired data is verified. After the verification is passed, the repaired data is integrated with the original usable data to form a new data set. This new integrated data set is set as the third data set to be tested, and the third data set to be tested is stored in a specified location on the server.

7. The method for identifying deviation from a clinical trial protocol according to claim 1, wherein: Step S105: Establish data detection rules, apply the data detection rules to the first data group to be detected, the second data group to be detected, and the third data group to be detected to obtain a data detection sequence, retrieve the detection models corresponding to the data detection sequence in the clinical trial protocol identification deviation model library, detect the first data group to be detected, the second data group to be detected, and the third data group to be detected according to the data detection sequence, and compare the detection results. According to the comparison results, generate a clinical trial protocol deviation result, including: According to the established data detection rules, the first to-be-detected data group, the second to-be-detected data group and the third to-be-detected data group are detected one by one; For the first data group to be tested, the second data group to be tested, and the third data group to be tested, corresponding detection models are retrieved from the clinical trial protocol identification deviation model library, and detection is performed according to a predetermined data detection order; After the detection is completed, the detection results corresponding to the first data group to be detected, the second data group to be detected, and the third data group to be detected are collected and recorded. The detection results include the identified potential deviation points, detailed information of the abnormal data, and the severity of the deviation. By comparison, common deviation points and deviation points in specific data groups are identified; Analyze the improvement of deviation identification on the repaired data in the third data group to be tested and the original data in the first data group to be tested and the second data group to be tested, and generate a clinical trial protocol deviation result report based on the comparison results, The clinical trial protocol deviation result report includes the deviation points identified in each group of data, including the type of deviation, the cause of the deviation, and the degree of impact of the deviation, as well as the impact of the repair on the accuracy of deviation identification. The generated deviation result report is output to the specified location.

8. The method for identifying deviation from a clinical trial protocol according to claim 7, wherein: Analyze the improvement of deviation identification on the repaired data in the third data group to be detected and the original data in the first data group to be detected and the second data group to be detected, including: Compare the number of deviation points in the first data group to be detected, the second data group to be detected and the third data group to be detected. If the number of deviation points of the repaired data in the third data group to be detected is reduced, the repair operation is effective; Comparing the repaired data in the third data group to be detected with the original data in the first data group to be detected and the second data group to be detected; By comparing the data before and after the repair, if the number of deviating points in the repaired data is reduced, the repair effect is positive; By comparing the recognition accuracy of deviation points in the data before and after repair, the improvement of deviation recognition accuracy caused by the repair operation can be evaluated. If the repaired data has a high accuracy in deviation recognition, the repair operation reduces the number of deviation points and improves the accuracy of deviation recognition in the clinical trial protocol.

9. A clinical trial protocol deviation identification and management system, using a clinical trial protocol deviation identification method as claimed in claims 1 to 8, characterized in that: It includes a server side, a backend control side and a user side. The server side establishes a communication connection with the backend control side and the user side. The server side includes: The server side obtains clinical trial data, which includes patient information data, clinical trial operation data, clinical observation data and other external data; The server creates corresponding data storage location labels for each data in the clinical trial data, stores each data in the clinical trial data in the corresponding data storage location, and simulates the use of data in each data storage location, creates labels for unusable data to obtain data to be repaired, and creates labels for usable data to obtain usable data; The server side builds a clinical trial protocol identification deviation model library, which includes prediction models, pattern recognition models, log analysis models, and behavior tracking models. It also builds data detection rules, and generates a data inspection sequence by passing the available data through the data detection rules. The server sets the usable data as the first data group to be detected, sets the data to be repaired and the usable data to form the second data group to be detected, repairs the data to be repaired to obtain repaired data, and sets the repaired data and the usable data to the third data group to be detected; The server side establishes data detection rules, applies the first data group to be detected, the second data group to be detected, and the third data group to be detected to the data detection rules, obtains the data detection sequence, retrieves the corresponding detection models in the data detection sequence in the clinical trial protocol identification deviation model library, detects the first data group to be detected, the second data group to be detected, and the third data to be detected according to the data detection sequence, and compares the detection results. Based on the comparison results, the clinical trial protocol deviation result is generated, and the server side sends the clinical trial protocol deviation result to the user side.

Citation Information

Patent Citations

  • Quality control method, device and equipment for medical data detection, medium and program product

    CN115691722A

  • Clinical test quality control method and equipment for scheme deviation semi-quantitative evaluation

    CN116864050A

  • Communication content monitoring method based on Internet of Things

    CN118473902A

Cited By

  • Clinical test quality evaluation method and system based on artificial intelligence

    CN119479958A