Self-adaptive clinical test monitoring method and device based on uncertainty prediction

Abstract

The invention discloses a self-adaptive clinical test monitoring method and device based on uncertainty prediction, and belongs to the technical field of data processing.The self-adaptive clinical test monitoring method based on uncertainty prediction comprises the steps that monitoring indexes and experiment conditions of clinical test monitoring are obtained; setting an uncertainty threshold value of a monitoring index according to experimental conditions and historical data; collecting test information at a first sampling frequency, and calculating real-time uncertainty by using a Bayesian model; if the real-time uncertainty exceeds the uncertainty threshold, the test information is collected at a second sampling frequency, the second sampling frequency being greater than the first sampling frequency. According to the method, through uncertainty prediction monitoring and feedback optimization, clinical test data can be dynamically, flexibly and efficiently monitored, through combination of uncertainty analysis, dynamic sampling and trend feedback, the monitoring frequency and threshold setting can be adaptively adjusted under the data fluctuation condition, and more reliable test data support is provided.

Description

Technical Field

[0001] The present application belongs to the field of data processing technology, and specifically relates to an adaptive clinical trial monitoring method and device based on uncertainty prediction. Background Art

[0002] With the rapid development of precision medicine and personalized treatment, clinical trials are becoming increasingly complex, particularly in terms of data structure and management. This complexity stems not only from the multidimensionality, diversity, and scale of data but also from the high-precision capture and analysis of dynamic changes and personalized characteristics. This trend is particularly evident in high-risk areas such as drug development and early dose safety testing, which place extremely high demands on data accuracy and real-time availability while also facing the challenge of rising data collection costs.

[0003] Traditional monitoring methods typically use a fixed sampling frequency to collect data. While this approach can meet the needs of routine trials to a certain extent, it is unable to cope with the more complex and dynamic clinical environment.

[0004] On the one hand, a fixed sampling frequency can lead to inefficient resource utilization. For example, when data fluctuations are minimal, fixed-frequency sampling can generate a large amount of redundant data, increasing storage and processing costs. On the other hand, the limitations of a fixed sampling frequency can lead to insufficient capture of critical data, especially when data exhibits sudden changes during the experiment. This omission of data can directly impact the reliability of the test results, exacerbating potential test risks. Summary of the Invention

[0005] The purpose of this application is to provide an adaptive clinical trial monitoring method and device based on uncertainty prediction to dynamically adjust the sampling frequency when uncertainty increases, realize real-time monitoring of high-risk data, and reduce data omissions.

[0006] According to a first aspect of an embodiment of the present application, a method for adaptive clinical trial monitoring based on uncertainty prediction is provided, which may include:

[0007] Obtain monitoring indicators and experimental conditions for clinical trial monitoring;

[0008] Set uncertainty thresholds for monitoring indicators based on experimental conditions and historical data;

[0009] Collecting test information at a first sampling frequency and calculating real-time uncertainty using a Bayesian model;

[0010] If the real-time uncertainty exceeds the uncertainty threshold, the test information is collected at a second sampling frequency, where the second sampling frequency is greater than the first sampling frequency.

[0011] In some optional embodiments of the present application, obtaining monitoring indicators and experimental conditions for clinical trial monitoring includes:

[0012] Obtain drug dosage safety, efficacy indicators, and experimental conditions monitored in clinical trials;

[0013] The experimental conditions included: high-dose group, medium-dose group and low-dose group.

[0014] In some optional embodiments of the present application, the uncertainty threshold of the monitoring indicator is set according to experimental conditions and in combination with historical data, including:

[0015] Use federated learning methods to obtain historical experimental data from multiple data sources;

[0016] The historical experimental data from multiple data sources are integrated to obtain historical data.

[0017] In some optional embodiments of the present application, after integrating historical experimental data from multiple data sources to obtain historical data, the following steps are further included:

[0018] Conduct risk assessment on historical data to obtain the risk level of historical data;

[0019] Perform data quality assessment on historical data according to the risk level to obtain the quality level of the historical data.

[0020] In some optional embodiments of the present application, before performing risk assessment on historical data to obtain the risk level of the historical data, the adaptive clinical trial monitoring method based on uncertainty prediction further includes:

[0021] Clean historical data to remove duplicate records, fill missing values and handle outliers;

[0022] Among them, statistical and machine learning methods are used to fill missing values and handle outliers;

[0023] Conduct risk assessment on historical data to obtain the risk level of historical data, including:

[0024] Feature engineering was performed on historical data to extract uncertainty values for dose levels, patient characteristics, trial phases, and historical data.

[0025] In some optional embodiments of the present application, the trial information includes: values of key indicators, timestamps, dosage levels and basic characteristics of the patient.

[0026] In some optional embodiments of the present application, if the current uncertainty exceeds the uncertainty threshold, after collecting the trial information at the second sampling frequency, the adaptive clinical trial monitoring method based on uncertainty prediction further includes:

[0027] Within the preset detection window, if the real-time uncertainty exceeds the uncertainty threshold for more than a preset number of times, test information is collected at a third sampling frequency and / or the uncertainty threshold is adjusted.

[0028] According to a second aspect of an embodiment of the present application, there is provided an adaptive clinical trial monitoring device based on uncertainty prediction, which may include:

[0029] Acquisition module, used to obtain monitoring indicators and experimental conditions for clinical trial monitoring;

[0030] The threshold setting module is used to set the uncertainty threshold of the monitoring indicator according to the experimental conditions and combined with historical data;

[0031] a calculation module, configured to collect test information at a first sampling frequency and calculate real-time uncertainty using a Bayesian model;

[0032] The threshold adjustment module is used to collect test information at a second sampling frequency when the real-time uncertainty exceeds the uncertainty threshold, and the second sampling frequency is greater than the first sampling frequency.

[0033] According to a third aspect of an embodiment of the present application, an electronic device is provided, which may include:

[0034] processor;

[0035] a memory for storing processor-executable instructions;

[0036] The processor is configured to execute instructions to implement the adaptive clinical trial monitoring method based on uncertainty prediction as shown in any one of the embodiments of the first aspect.

[0037] According to the fourth aspect of the embodiments of the present application, a storage medium is provided. When the instructions in the storage medium are executed by the processor of an information processing device or a server, the information processing device or the server implements an adaptive clinical trial monitoring method based on uncertainty prediction as shown in any one of the embodiments of the first aspect.

[0038] The above technical solution of this application has the following beneficial technical effects:

[0039] The method in this application embodiment enables dynamic, flexible, and efficient monitoring of clinical trial data through monitoring and feedback optimization based on uncertainty prediction. By combining uncertainty analysis, dynamic sampling, and trend feedback, it enables adaptive adjustment of monitoring frequency and threshold settings under data fluctuations, providing more reliable test data support. This provides important guarantees for the stability and scientific nature of test data while reducing resource waste and experimental interference caused by frequent false alarms. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a flow chart of an adaptive clinical trial monitoring method based on uncertainty prediction in an exemplary embodiment of the present application;

[0041] Figure 2 is a schematic structural diagram of an adaptive clinical trial monitoring device based on uncertainty prediction in an exemplary embodiment of the present application;

[0042] Figure 3 is a schematic structural diagram of an electronic device in an exemplary embodiment of the present application;

[0043] Figure 4 It is a schematic diagram of the hardware structure of an electronic device in an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0044] To make the objectives, technical solutions, and advantages of this application more clearly understood, this application is further described below in conjunction with specific embodiments and with reference to the accompanying drawings. It should be understood that these descriptions are merely illustrative and are not intended to limit the scope of this application. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion in the concepts of this application.

[0045] The accompanying drawings illustrate schematic diagrams of layer structures according to embodiments of the present application. These figures are not drawn to scale; for clarity, some details are exaggerated and some details may be omitted. The shapes of the various regions and layers shown in the figures, as well as their relative sizes and positional relationships, are merely exemplary and may deviate in practice due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art may design regions / layers with different shapes, sizes, and relative positions as needed.

[0046] Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0047] In the description of this application, it should be noted that the terms "first", "second" and "third" are used for descriptive purposes only and should not be understood as indicating or implying relative importance.

[0048] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0049] The adaptive clinical trial monitoring method based on uncertainty prediction provided by the embodiment of the present application is described in detail below with reference to the accompanying drawings through specific embodiments and their application scenarios.

[0050] like Figure 1 As shown, in a first aspect of an embodiment of the present application, a method for adaptive clinical trial monitoring based on uncertainty prediction is provided, which may include:

[0051] S110: Obtain monitoring indicators and experimental conditions for clinical trial monitoring;

[0052] S120: Setting uncertainty thresholds for monitoring indicators based on experimental conditions and combined with historical data;

[0053] S130: Collecting test information at a first sampling frequency and calculating real-time uncertainty using a Bayesian model;

[0054] S140: If the real-time uncertainty exceeds the uncertainty threshold, collect test information at a second sampling frequency, where the second sampling frequency is greater than the first sampling frequency.

[0055] Through monitoring and feedback optimization based on uncertainty prediction, the method in this embodiment enables dynamic, flexible, and efficient monitoring of clinical trial data. By combining uncertainty analysis, dynamic sampling, and trend feedback, it enables adaptive adjustment of monitoring frequency and threshold settings under data fluctuations, providing more reliable test data support. This provides important guarantees for the stability and scientific nature of test data while reducing resource waste and experimental interference caused by frequent false alarms.

[0056] The method of this embodiment is applied to an adaptive clinical trial monitoring system based on uncertainty prediction, which may include: an acquisition module, a threshold setting module, a calculation module, and a threshold adjustment module. In this embodiment, the monitoring indicators may include safety and efficacy indicators of drug dosage; the experimental conditions may include high-dose group, medium-dose group, and low-dose group, and an uncertainty threshold U_thresh is set for each indicator. This is a range within which data fluctuations are allowed, which can help the team judge the stability and rationality of the data. The system will analyze historical data, combine industry literature, and compare data fluctuations in similar trials to generate a suitable U_thresh for each key indicator.

[0057] In a specific embodiment, identification monitoring indicators can be determined by determining the key variables of the test, the safety (SafetyParameter) and efficacy (EfficacyParameter) of the drug dosage. These variables will affect the main results of the test and therefore require continuous monitoring. Thresholds are then set for different experimental conditions. Here, the dose groups can be divided (low-dose group, medium-dose group, high-dose group), and a lower U_thresh is set for the high-dose group to strictly control the risk. At the same time, a higher U_thresh is set for the low-dose group so that it can tolerate more fluctuations. Next, an initial threshold is constructed. Here, based on previous test data, the U_thresh under each condition can be constructed to capture the reasonable fluctuation range of the data and avoid frequent data false alarms after the start of the test.

[0058] For example, in a vaccine dosage trial, the U_thresh setting for the high-dose group is lower, and the system will be more sensitive to detecting slight fluctuations in the data, such as changes in blood concentration, so that even small abnormalities can be captured in time; for the low-dose group, the U_thresh setting is higher, and the system allows more fluctuations to avoid low-risk data causing too many false alarms.

[0059] By dynamically setting U_thresh, the system can flexibly adjust the fluctuation range according to the dose and patient characteristics, so that the stability of the data can be adaptively optimized according to different conditions in the early stages of the trial, laying a good foundation for subsequent data monitoring.

[0060] In some embodiments, obtaining monitoring indicators and experimental conditions for clinical trial monitoring includes:

[0061] Obtain drug dosage safety, efficacy indicators, and experimental conditions monitored in clinical trials;

[0062] The experimental conditions included: high-dose group, medium-dose group and low-dose group.

[0063] During the initial stages of trial design, the core indicators to be monitored (such as drug dosage safety and efficacy) are clearly defined. An uncertainty threshold (U_thresh) is set for each indicator based on the different experimental conditions (high-dose, medium-dose, and low-dose groups). This allows for fluctuations in the data and helps determine its stability and rationality. A machine learning model analyzes historical data to generate an appropriate U_thresh for each key indicator.

[0064] In some embodiments, setting uncertainty thresholds for monitoring indicators based on experimental conditions and in combination with historical data includes:

[0065] Use federated learning methods to obtain historical experimental data from multiple data sources;

[0066] The historical experimental data from multiple data sources are integrated to obtain historical data.

[0067] In this embodiment, historical experimental data can be obtained from multiple data sources, including internal databases, external public databases (PubMed, ClinicalTrials.gov), and databases of partner institutions. Data integration technology is used to integrate data from different data sources to ensure data integrity and consistency. Federated learning technology is used to integrate data from different data sources through model aggregation without directly sharing data. This approach not only protects data privacy but also improves the efficiency and security of data integration.

[0068] Specifically, obtaining historical experimental data based on federated learning can include:

[0069] Decentralized model updates: Traditional federated learning relies on centralized servers for model aggregation, which poses single points of failure and trust risks. We use blockchain to build a decentralized model update platform. Model updates from all participating institutions are recorded on the blockchain, ensuring data immutability, transparency, and traceability.

[0070] Smart Contract Automation: Leverage smart contracts to automatically manage model updates and aggregation processes. Smart contracts define the rules and incentives for model updates, ensuring that participants contribute to model updates in accordance with the agreement, improving the automation and fairness of the system.

[0071] Differential Privacy Protection:

[0072] Enhanced data privacy: Differential privacy noise is added to local model updates to prevent sensitive information from being leaked through model parameters. Even if model updates are obtained by other nodes, the original data cannot be inferred, further protecting data privacy.

[0073] Adaptive adjustment of privacy parameters: Based on the importance and sensitivity of the data, smart contracts can adaptively adjust the parameters of differential privacy (such as privacy budget) to find the best balance between privacy protection and model accuracy.

[0074] Integration and optimization of federated learning and blockchain:

[0075] Consensus mechanism optimization: The adoption of lightweight consensus mechanisms designed specifically for federated learning, such as the validator-based Byzantine fault-tolerant algorithm, reduces the communication overhead and latency of the blockchain network and improves the efficiency of model aggregation.

[0076] Model update verification: Model update verification nodes are introduced into the blockchain network, and cryptographic technologies such as zero-knowledge proof are used to verify whether the submitted model updates meet expectations, thereby improving the security and reliability of the system.

[0077] In a specific embodiment, the process of acquiring historical experimental data using federated learning includes:

[0078] Step 1: Local training and differential privacy processing: Each participating institution trains the model locally using its own data. Differential privacy noise is added to the model update to generate privacy-preserving model parameters.

[0079] Step 2: Submit the model update and record it on the blockchain. Submit the encrypted model update to the blockchain network. The blockchain verifies the legitimacy of the model update through a consensus mechanism and records it on the chain.

[0080] Step 3: Global model aggregation and distribution: The smart contract automatically executes the model aggregation algorithm to generate a new global model. The new global model is distributed to all participating institutions through the blockchain network.

[0081] Combining differential privacy and blockchain technology, it doubles the privacy and security of data and model updates.

[0082] Improved system efficiency and reliability: A decentralized architecture eliminates single points of failure, and an optimized consensus mechanism accelerates model aggregation. The integration of blockchain, differential privacy, and federated learning creates a novel data integration and model training framework, a first for multi-source clinical trial data integration.

[0083] This innovative federated learning mechanism is particularly well-suited for the medical field, where data privacy and security are paramount. By ensuring that participating institutions' data remain within their domain and that model updates are secure and reliable, it significantly enhances the willingness and efficiency of multi-institutional collaboration, providing strong support for high-quality clinical trial data analysis.

[0084] In some embodiments, after integrating historical experimental data from multiple data sources to obtain historical data, the method further includes:

[0085] Conduct risk assessment on historical data to obtain the risk level of historical data;

[0086] Perform data quality assessment on historical data according to the risk level to obtain the quality level of the historical data.

[0087] In this embodiment, an adaptive data quality assessment algorithm can be introduced to dynamically adjust the assessment criteria based on data characteristics. For example, for high-risk test data, a stricter assessment criteria can be used; for low-risk test data, a relatively looser assessment criteria can be used.

[0088] In order to improve the efficiency and accuracy of data quality assessment, this embodiment proposes an adaptive data quality assessment method based on risk level, which dynamically adjusts the data quality assessment standard according to the risk level of the test data, thereby optimizing resource utilization while ensuring data reliability.

[0089] Specifically, the evaluation method may include:

[0090] Define risk levels: Classify trial data into high, medium, or low risk levels based on their impact and sensitivity to trial results. For example, data related to key efficacy indicators is defined as high risk, while general background data may be defined as low risk.

[0091] Risk factor determination: Based on the source, type and purpose of the data, corresponding risk factors are set to assess the risk level of the data.

[0092] Adaptive assessment criteria adjustment:

[0093] Dynamically adjust assessment standards: Automatically apply different levels of quality assessment standards based on the risk level of the data. Stricter quality assessment standards, such as stricter missing value thresholds and outlier detection rules, are applied to high-risk data, while less stringent standards are applied to low-risk data.

[0094] Resource optimization: By focusing on high-risk data and rationally allocating resources for data cleaning and quality control, the overall data processing efficiency can be improved.

[0095] More specifically, the evaluation method may include:

[0096] Establish a rule base: Create a rule base that contains assessment criteria corresponding to different risk levels. The rule base includes specific treatment standards for various data quality issues (such as missing values, duplicate values, and outliers).

[0097] Automated processes: Integrate automated processes for risk level identification and standard application into data quality assessment tools to ensure an efficient and accurate assessment process.

[0098] Feedback mechanism:

[0099] Quality assessment feedback: During the data quality assessment process, the system will generate an assessment report, pointing out the quality issues in the data at each risk level for review by data managers.

[0100] Continuous improvement: Based on assessment results and user feedback, we continuously improve risk classification and assessment standards to enhance the applicability and effectiveness of assessment methods.

[0101] This method is simple and easy to implement, requiring no complex algorithms or models. It ensures the quality of critical data and the reliability of test results through strict control of high-risk data. By applying more relaxed standards to low-risk data, data processing time and costs can be reduced, improving overall efficiency.

[0102] For example, in a clinical trial, a patient's biomarker data is defined as high-risk, while their general demographic information is defined as low-risk. This approach allows the system to rigorously assess the quality of biomarker data to ensure the accuracy of this critical data. Simultaneously, the assessment criteria for demographic information are appropriately relaxed to reduce unnecessary resource investment.

[0103] This embodiment combines data risk levels with quality assessment criteria, proposing an adaptive data quality assessment method that dynamically adjusts assessment criteria based on data risk. This method improves resource utilization efficiency while ensuring critical data quality, demonstrating significant practical value and innovation.

[0104] In some embodiments, before performing risk assessment on historical data to obtain a risk level of the historical data, the adaptive clinical trial monitoring method based on uncertainty prediction further includes:

[0105] Clean historical data to remove duplicate records, fill missing values and handle outliers;

[0106] Among them, statistical and machine learning methods are used to fill missing values and handle outliers;

[0107] Conduct risk assessment on historical data to obtain the risk level of historical data, including:

[0108] Feature engineering was performed on historical data to extract uncertainty values for dose levels, patient characteristics, trial phases, and historical data.

[0109] The statistical method of this embodiment can use mean filling and median filling, and the machine learning method can use random forest filling. Use feature selection method and feature extraction method to optimize the feature set.

[0110] In some embodiments, the trial information includes: values of key indicators, timestamps, dosage levels, and basic characteristics of the patient.

[0111] For example, a suitable machine learning model is selected to predict the uncertainty threshold U_thresh for different dose groups. (Adaptive model selection: An adaptive model selection algorithm is used to dynamically select the optimal model based on data characteristics and trial stage. In the early stages of the trial, a simple model (linear regression) is selected for rapid prediction; in the later stages of the trial, a complex model (XGBoost) is selected for accurate prediction. The preprocessed historical data is divided into a training set, a validation set, and a test set. The training set is used to train the machine learning model, and the validation set is used for hyperparameter tuning. Cross-validation and grid search are used to ensure the generalization ability and prediction accuracy of the model.

[0112] In some embodiments, after collecting the trial information at the second sampling frequency if the current uncertainty exceeds the uncertainty threshold, the adaptive clinical trial monitoring method based on uncertainty prediction further includes:

[0113] Within the preset detection window, if the real-time uncertainty exceeds the uncertainty threshold for more than a preset number of times, test information is collected at a third sampling frequency and / or the uncertainty threshold is adjusted.

[0114] After the trial is initiated, data collection begins at a preset sampling frequency (e.g., every four hours). Each data point collected includes the values of key indicators (e.g., blood concentrations), a timestamp, the dose level, and basic patient characteristics. Each collected data point, D_t, serves as the foundational information for the trial. A Bayesian model is used to analyze the data in real time, assessing data point volatility and calculating the current uncertainty, U_t, to determine whether the sampling frequency needs to be adjusted.

[0115] Specifically, the experiment may include:

[0116] Data collection: Data is automatically collected according to the planned sampling frequency. Each recorded data includes the key indicator value (such as the blood concentration value at a certain dose), the collection time TimeStamp and the patient's relevant characteristics.

[0117] Real-time data analysis: Data is analyzed instantly through the Bayesian model to update the posterior distribution of the test data so that changes in data points can be reflected in the model in a timely manner.

[0118] Uncertainty calculation: Calculate the uncertainty U_t of the collected data and compare it with the set U_thresh to determine whether the data exceeds the normal fluctuation range.

[0119] In personalized medicine trials, physiological indicators can vary significantly between patients. For example, heart rate and blood pressure can be affected by multiple factors. After collecting a patient's heart rate data, we analyze and compare it with data from other patients to calculate U_t. If U_t is greater than U_thresh, this indicates that the data fluctuations are outside the normal range, suggesting the need for closer monitoring of the patient's heart rate or increasing sampling frequency to gain a more detailed understanding of their condition.

[0120] The method of this embodiment calculates the uncertainty U_t in real time based on the data collected each time, quickly responds to data fluctuations, and can promptly identify sudden data fluctuations caused by individual differences or dosage changes during the test process, providing a data basis for subsequent dynamic sampling adjustments.

[0121] When a data point's U_t exceeds the set uncertainty threshold U_thresh, the system automatically enters an adaptive sampling adjustment process to increase sampling density, ensuring more data can be collected intensively despite large data fluctuations. If multiple consecutive high-uncertainty data points are detected, an automatic alert is generated, prompting the test team to reassess the sampling strategy or test conditions to ensure the stability of the test data.

[0122] Adjust the sampling frequency: When U_t exceeds U_thresh, the sampling frequency is automatically increased, for example, from 4 hours to 2 hours, to ensure that data can be collected more intensively.

[0123] Monitoring data continuity: The system tracks the persistence of high uncertainty data points and generates an alert if U_thresh is exceeded multiple times in a set time.

[0124] Generate an early warning notification: If U_t of n consecutive data points exceeds U_thresh, an early warning is triggered, reminding the experimental team to review the experimental design or further adjust the sampling density.

[0125] In drug safety trials, when safety indicators of drug dosage (such as blood concentrations) show fluctuations and U_t continuously exceeds U_thresh, the system will adjust the sampling frequency from every 4 hours to every 2 hours to ensure that a richer data sample is obtained within the data interval with high uncertainty. If high volatility is detected repeatedly, an automatic warning will be triggered, reminding the research team to check whether the dosage level needs to be adjusted or the patient group needs to be reallocated to reduce potential risks.

[0126] Adaptive sampling adjustment automatically increases sampling frequency based on uncertainty, ensuring more accurate recording of critical data even in the face of significant data fluctuations. Furthermore, an early warning mechanism helps the test team identify potential safety hazards in advance, ensuring test stability and data integrity.

[0127] Within a set monitoring window T (e.g., 24 hours), uncertainty fluctuation analysis is regularly performed on all data points to identify overall trends and generate feedback reports. If the majority of data points U_t within the T window exceed U_thresh, U_thresh or the sampling frequency is automatically adjusted to accommodate data fluctuations, ensuring a more stable subsequent test data collection process.

[0128] Trend analysis: Calculate the fluctuation of the data point U_t within the window T and evaluate the overall trend of uncertainty to determine whether the data fluctuation is increasing or decreasing.

[0129] Threshold feedback adjustment: If the trend analysis results show that data fluctuations are increasing, moderately relax U_thresh as needed, or increase the sampling frequency to ensure continuous monitoring of highly volatile data.

[0130] Automatic optimization: Based on feedback data, the sampling and monitoring strategies are dynamically optimized, and the initial thresholds and sampling frequencies are gradually restored when fluctuations slow down, maintaining a reasonable allocation of monitoring resources.

[0131] To further improve the ability to respond to data fluctuations, a trend relaxation factor, TrendFactor, is introduced. This factor dynamically adjusts U_thresh based on the average uncertainty level within the time window T, allowing the threshold to be automatically relaxed to adapt to changes when fluctuations increase.

[0132] The formula is as follows:

[0133]

[0134] U thresh，new : is the adjusted uncertainty threshold.

[0135] γ: Trend response coefficient, used to control the degree of relaxation.

[0136] ∑U t : is the sum of the uncertainties of all data points in window T.

[0137] n: is the total number of data points in the window.

[0138] This formula takes the ratio of the average uncertainty level in the window to the threshold as the increment, so that U_thresh is automatically relaxed according to the actual fluctuation, avoiding frequent false alarms caused by slight fluctuations.

[0139] In vaccine efficacy trials, if antibody concentration U_t fluctuates significantly and persistently above U_thresh within a 24-hour window, U_thresh is automatically relaxed to accommodate the new fluctuation. Dynamically adjusting U_thresh automatically adapts to volatile efficacy data, ensuring consistent data collection while avoiding frequent false alarms caused by overly stringent threshold settings.

[0140] Throughout the test, all sampling frequency adjustments, uncertainty changes, and warning notifications are recorded in the database and updated in real time on the dashboard, allowing the test team to monitor the dynamic changes of the test at any time. After the test, a data fluctuation trend chart and adjustment records are automatically generated for the team to review and reference for subsequent test design.

[0141] Data storage: Real-time recording of each data point’s U_t, sampling frequency f_t, and alert status AlertStatus to provide complete test data.

[0142] Visualization: Uncertainty trends and sampling frequency adjustments are displayed on the dashboard, allowing the test team to observe data fluctuations and system responses in real time.

[0143] Generate trend charts: After the test, generate data fluctuation trend charts and uncertainty adjustment records to support the team in summarizing test data fluctuation patterns and analyzing key fluctuation points.

[0144] During vaccine trials, the team can use a dashboard to view real-time fluctuations in U_t and adjustments to the sampling frequency. After the trial, the generated trend charts and adjustment records help the research team analyze fluctuations during the trial and provide reference data support for optimizing sampling strategies for the next trial.

[0145] It should be noted that the method for adaptive clinical trial monitoring based on uncertainty prediction provided in the embodiments of the present application can be performed by an adaptive clinical trial monitoring device based on uncertainty prediction, or a control module in the adaptive clinical trial monitoring device based on uncertainty prediction for performing the method for adaptive clinical trial monitoring based on uncertainty prediction. In the embodiments of the present application, the method for adaptive clinical trial monitoring based on uncertainty prediction performed by an adaptive clinical trial monitoring device based on uncertainty prediction is used as an example to illustrate the device for adaptive clinical trial monitoring based on uncertainty prediction provided in the embodiments of the present application.

[0146] like Figure 2 As shown, in a second aspect of an embodiment of the present application, an adaptive clinical trial monitoring device based on uncertainty prediction is provided, which may include:

[0147] Acquisition module 210, for acquiring monitoring indicators and experimental conditions for clinical trial monitoring;

[0148] A threshold setting module 220 is used to set the uncertainty threshold of the monitoring indicator according to the experimental conditions and in combination with historical data;

[0149] a calculation module 230 for collecting test information at a first sampling frequency and calculating real-time uncertainty using a Bayesian model;

[0150] The threshold adjustment module 240 is configured to collect test information at a second sampling frequency when the real-time uncertainty exceeds the uncertainty threshold, where the second sampling frequency is greater than the first sampling frequency.

[0151] The adaptive clinical trial monitoring device based on uncertainty prediction in the embodiment of the present application can be a device, or a component, integrated circuit, or chip in a terminal. The device can be a mobile electronic device or a non-mobile electronic device. For example, the mobile electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, an in-vehicle electronic device, a wearable device, an ultra-mobile personal computer (UMPC), a netbook or a personal digital assistant (PDA), etc., and the non-mobile electronic device can be a server, a network attached storage (NAS), a personal computer (PC), a television (TV), an ATM or a self-service machine, etc., which is not specifically limited in the embodiment of the present application.

[0152] The adaptive clinical trial monitoring device based on uncertainty prediction in the embodiments of the present application can be a device having an operating system. The operating system can be an Android operating system, an iOS operating system, or other possible operating systems, which are not specifically limited in the embodiments of the present application.

[0153] The adaptive clinical trial monitoring device based on uncertainty prediction provided by the embodiment of the present application can achieve Figure 1 To avoid repetition, the various processes implemented in the method embodiment are not described here.

[0154] Alternatively, as Figure 3As shown, an embodiment of the present application also provides an electronic device 300, including a first processor 301, a first memory 302, and a program or instruction stored in the first memory 302 and executable on the first processor 301. When the program or instruction is executed by the first processor 301, each process of the above-mentioned embodiment of the adaptive clinical trial monitoring method based on uncertainty prediction is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0155] It should be noted that the electronic devices in the embodiments of the present application include the mobile electronic devices and non-mobile electronic devices mentioned above.

[0156] Figure 4 A schematic diagram of the hardware structure of an electronic device implementing an embodiment of the present application.

[0157] The hardware structure 400 of the electronic device includes but is not limited to: a radio frequency unit 401, a network module 402, an audio output unit 403, an input unit 404, a sensor 405, a display unit 406, a user input unit 407, an interface unit 408, a second memory 409, and a second processor 410.

[0158] Those skilled in the art will understand that the hardware structure 400 of the electronic device may also include a power source (such as a battery) to power each component, and the power source may be logically connected to the second processor 410 through a power management system, thereby implementing functions such as charging, discharging, and power consumption management through the power management system. Figure 4 The electronic device structure shown in the figure does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently, which will not be repeated here.

[0159] It should be understood that in the embodiment of the present application, the input unit 404 may include a graphics processing unit (GPU) 4041 and a microphone 4042. The graphics processor 4041 processes image data of still images or videos obtained by an image capture device (such as a camera) in video capture mode or image capture mode. The display unit 406 may include a display panel 4061, which may be configured in the form of a liquid crystal display, an organic light-emitting diode, etc. The user input unit 407 includes a touch panel 4071 and other input devices 4072. The touch panel 4071 is also called a touch screen. The touch panel 4071 may include two parts: a touch detection device and a touch controller. Other input devices 4072 may include, but are not limited to, a physical keyboard, function keys (such as volume control keys, power keys, etc.), a trackball, a mouse, and a joystick, which will not be described in detail here. The second memory 409 can be used to store software programs and various data, including but not limited to applications and operating systems. The second processor 410 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and applications, etc., and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the second processor 410 .

[0160] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of the above-mentioned embodiment of the adaptive clinical trial monitoring method based on uncertainty prediction are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0161] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer-readable storage medium, such as a computer read-only memory (ROM), random access memory (RAM), a magnetic disk, or an optical disk.

[0162] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned embodiment of the adaptive clinical trial monitoring method based on uncertainty prediction, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0163] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0164] It should be noted that, in this article, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element. In addition, it should be noted that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in the opposite order according to the functions involved. For example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0165] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), including a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0166] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose of this application and the scope of protection of the claims, all of which are within the protection of this application.

Claims

1. An adaptive clinical trial monitoring method based on uncertainty prediction, characterized in that: include: Obtain monitoring indicators and experimental conditions for clinical trial monitoring; Setting an uncertainty threshold of the monitoring indicator according to the experimental conditions and in combination with historical data; Collecting test information at a first sampling frequency and calculating real-time uncertainty using a Bayesian model; If the real-time uncertainty exceeds the uncertainty threshold, test information is collected at a second sampling frequency, where the second sampling frequency is greater than the first sampling frequency.

2. The adaptive clinical trial monitoring method based on uncertainty prediction according to claim 1, characterized in that: The monitoring indicators and experimental conditions for clinical trial monitoring include: Obtain drug dosage safety, efficacy indicators, and experimental conditions monitored in clinical trials; The experimental conditions include: a high-dose group, a medium-dose group and a low-dose group.

3. The adaptive clinical trial monitoring method based on uncertainty prediction according to claim 1, characterized in that: The step of setting the uncertainty threshold of the monitoring indicator according to the experimental conditions and in combination with historical data includes: Use federated learning methods to obtain historical experimental data from multiple data sources; The historical experimental data from multiple data sources are integrated to obtain historical data.

4. The adaptive clinical trial monitoring method based on uncertainty prediction according to claim 3, characterized in that: After integrating the historical experimental data from multiple data sources to obtain historical data, the method further includes: Performing a risk assessment on the historical data to obtain a risk level of the historical data; A data quality assessment is performed on the historical data according to the risk level to obtain a quality level of the historical data.

5. The adaptive clinical trial monitoring method based on uncertainty prediction according to claim 4, characterized in that: Before performing risk assessment on the historical data to obtain the risk level of the historical data, the adaptive clinical trial monitoring method based on uncertainty prediction further includes: Cleaning the historical data to remove duplicate records, fill missing values and process outliers; Among them, statistical and machine learning methods are used to fill missing values and handle outliers; The performing risk assessment on the historical data to obtain the risk level of the historical data includes: Feature engineering is performed on the historical data to extract uncertainty values for dose levels, patient characteristics, trial phases, and historical data.

6. The adaptive clinical trial monitoring method based on uncertainty prediction according to claim 1, characterized in that: The trial information includes: values of key indicators, timestamps, dosage levels and basic characteristics of patients.

7. The adaptive clinical trial monitoring method based on uncertainty prediction according to claim 1, characterized in that: After collecting the test information at the second sampling frequency if the current uncertainty exceeds the uncertainty threshold, the adaptive clinical trial monitoring method based on uncertainty prediction further includes: Within a preset detection window, if the real-time uncertainty exceeds the uncertainty threshold for more than a preset number of times, test information is collected at a third sampling frequency and / or the uncertainty threshold is adjusted.

8. An adaptive clinical trial monitoring device based on uncertainty prediction, characterized in that: include: Acquisition module, used to obtain monitoring indicators and experimental conditions for clinical trial monitoring; A threshold setting module, configured to set an uncertainty threshold of the monitoring indicator according to the experimental conditions and in combination with historical data; a calculation module, configured to collect test information at a first sampling frequency and calculate real-time uncertainty using a Bayesian model; A threshold adjustment module is configured to collect test information at a second sampling frequency when the real-time uncertainty exceeds the uncertainty threshold, where the second sampling frequency is greater than the first sampling frequency.

9. An electronic device, characterized in that: include: A processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the steps of the adaptive clinical trial monitoring method based on uncertainty prediction as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium stores a program or instruction, and when the program or instruction is executed by the processor, the steps of the adaptive clinical trial monitoring method based on uncertainty prediction as described in any one of claims 1 to 7 are implemented.

Images