Drug risk monitoring method and device and electronic equipment

Abstract

The invention discloses a drug risk monitoring method and device and electronic equipment. The method comprises the following steps: obtaining individual feature information and initial medication data of a target object; a risk assessment model is adopted to analyze the individual feature information and the initial medication data to obtain a risk assessment result, an early warning strategy is generated according to the risk assessment result, the risk assessment result is used for reflecting the risk degree of the initial medication data, the risk assessment model is trained by adopting data sources in multiple dimensions, and the early warning strategy is generated according to the risk assessment model. The feedback condition of the initial medication data is optimized according to the target object; and determining target medication data of the target object according to the early warning strategy. According to the invention, the problems that the timeliness and accuracy of risk early warning are insufficient because most drug risk monitoring methods in related technologies depend on a single data source and a risk assessment model lacks real-time performance and dynamic adjustment capability are solved; and the real risk condition of the medicine in the wide crowd and diversified use scenes cannot be comprehensively captured.

Description

Technical Field

[0001] The present application relates to the technical field of drug risk monitoring, and in particular, to a drug risk monitoring method, device and electronic equipment. Background Art

[0002] Drug risk monitoring is an important part of ensuring public health and safety. It covers the risk tracking, identification, analysis and evaluation from drug research and development, production, sales to post-market use. This process aims to timely discover and evaluate the potential harm that drugs may cause to patients, such as adverse reactions, side effects and drug interactions, in order to ensure drug safety and improve drug safety.

[0003] However, when faced with the limitations of data sources, drug risk monitoring technologies in related technologies often fail to capture the risk information of drugs in a wider range of populations and more diverse usage scenarios, resulting in biased risk assessment results and failure to accurately reflect the true risk status of drugs. In addition, the real-time and dynamic adjustment capabilities of drug risk monitoring systems are also relatively limited, focusing on static or periodic risk assessments, and failing to analyze real-time data streams and dynamically adjust risk assessment models and early warning strategies based on new data and feedback, reducing the timeliness and accuracy of risk warnings.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present application provide a drug risk monitoring method, device and electronic device to at least solve the technical problem that the drug risk monitoring methods in related technologies mostly rely on a single data source, and the risk assessment model lacks real-time and dynamic adjustment capabilities, resulting in insufficient timeliness and accuracy of risk warnings, and cannot fully capture the actual risk situation of drugs in a wide range of populations and diverse usage scenarios.

[0006] According to one aspect of an embodiment of the present application, a drug risk monitoring method is provided, including: obtaining individual characteristic information and initial medication data of a target object; using a risk assessment model to analyze the individual characteristic information and the initial medication data to obtain a risk assessment result, and generating an early warning strategy based on the risk assessment result, wherein the risk assessment result is used to reflect the risk level of the initial medication data, the risk assessment model is trained using data sources in multiple dimensions, and is optimized based on the feedback of the target object on the initial medication data; and determining the target medication data of the target object based on the early warning strategy.

[0007] Optionally, the risk assessment model is trained in the following manner: obtaining historical drug data; extracting historical drug usage information and historical patient feedback information from the historical drug data; determining a multidimensional feature data set based on the historical drug usage information and historical patient feedback information, and dividing the multidimensional feature data set into a training set and a test set; determining an initial model, and training the initial model based on the training set, and optimizing the initial model based on the test set to obtain a risk assessment model.

[0008] Optionally, obtaining historical drug data includes: obtaining first drug data under multiple dimensions, wherein the first drug data is used to represent a single drug data corresponding to each dimension; determining a feature weight value corresponding to the first drug data, wherein the feature weight value is used to represent a contribution of the first drug data to drug risk assessment; comparing the feature weight value with a preset weight threshold, and determining the first drug data corresponding to the feature weight value exceeding the preset weight threshold as the second drug data; and fusing the second drug data to obtain historical drug data.

[0009] Optionally, a multi-dimensional feature data set is determined based on historical drug usage information and historical patient feedback information, including: extracting core features from the historical drug usage information and historical patient feedback information, wherein the core features are used to represent data features related to drug risk assessment; normalizing the core features; and fusing the normalized core features to obtain a multi-dimensional feature data set.

[0010] Optionally, the initial model is optimized based on the test set, including: using the initial model to process the test set to obtain historical risk prediction results; comparing the historical risk prediction results with actual historical risk events to obtain comparison results; when the comparison result indicates that the deviation value between the historical risk prediction results and the actual historical risk events is higher than a first preset threshold, optimizing the initial model.

[0011] Optionally, the method also includes: obtaining real-time data of the target object after using the target medication data, wherein the real-time data includes real-time feedback information, real-time drug usage information and adverse reaction reports of the target object; optimizing the risk assessment model based on the real-time data, updating the data sources in multiple dimensions based on the real-time data, and improving the user interface where the target object is located based on the real-time data.

[0012] Optionally, optimizing the risk assessment model based on real-time data includes: performing incremental learning on the risk assessment model using a gradient boosting decision tree based on the real-time data; determining performance indicators of the risk assessment model after performing incremental learning, wherein the performance indicators include accuracy, recall and F1 score indicators of the risk assessment model; and updating model parameters of the risk assessment model when the accuracy is lower than a second preset threshold, the recall is lower than a third preset threshold and the F1 score indicator is lower than a fourth preset threshold.

[0013] According to another aspect of an embodiment of the present application, a drug risk monitoring device is also provided, including: an acquisition module, used to acquire individual characteristic information and initial medication data of a target object; an analysis module, used to analyze the individual characteristic information and initial medication data using a risk assessment model to obtain a risk assessment result, and generate an early warning strategy based on the risk assessment result, wherein the risk assessment result is used to reflect the risk level of the initial medication data, the risk assessment model is trained using data sources in multiple dimensions, and is optimized based on the feedback of the target object on the initial medication data; an early warning module, used to determine the target medication data of the target object based on the early warning strategy.

[0014] According to another aspect of an embodiment of the present application, there is also provided an electronic device, comprising: a memory and a processor, wherein the memory is used to store program instructions; and the processor is connected to the memory and is used to execute the above-mentioned drug risk monitoring method.

[0015] According to another aspect of the embodiment of the present application, a non-volatile storage medium is also provided, the non-volatile storage medium includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the above-mentioned drug risk monitoring method by running the computer program.

[0016] According to another aspect of the embodiments of the present application, a computer program product is also provided, including computer instructions, which implement the above-mentioned drug risk monitoring method when executed by a processor.

[0017] In an embodiment of the present application, individual characteristic information and initial medication data of a target object are obtained; a risk assessment model is used to analyze the individual characteristic information and initial medication data to obtain a risk assessment result, and an early warning strategy is generated based on the risk assessment result, wherein the risk assessment result is used to reflect the risk level of the initial medication data, the risk assessment model is trained using data sources in multiple dimensions, and is optimized based on the feedback of the target object on the initial medication data; the target medication data of the target object is determined based on the early warning strategy, thereby achieving the purpose of accurately predicting and dynamically preventing and controlling drug risks, thereby achieving the technical effect of improving the real-time, accuracy and personalization level of drug risk monitoring and early warning, and further solving the technical problem that the drug risk monitoring methods in related technologies mostly rely on a single data source, and the risk assessment model lacks real-time and dynamic adjustment capabilities, resulting in insufficient timeliness and accuracy of risk warnings, and the inability to fully capture the true risk situation of drugs in a wide range of populations and in diverse usage scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0019] Figure 1 It is a hardware structure diagram of a computer terminal for implementing a drug risk monitoring method according to an embodiment of the present application;

[0020] Figure 2 is a flow chart of a drug risk monitoring method according to an embodiment of the present application;

[0021] Figure 3 It is a structural diagram of a drug risk monitoring device according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] First, some nouns or terms that appear in the process of explaining the embodiments of the present application are subject to the following explanations:

[0025] GBDT (Gradient Boosting Decision Tree): A machine learning algorithm that belongs to the ensemble learning method. It optimizes the prediction model by iteratively adding decision trees, and each tree is trained on the residual of the previous tree, that is, the prediction error. GBDT performs well in handling classification and regression problems, can handle large amounts of data and automatically perform feature selection.

[0026] TensorFlow: An open source machine learning framework that provides tools for building and training various machine learning models, especially deep learning models. TensorFlow supports distributed computing, which can accelerate the model training process.

[0027] PyTorch: An open source machine learning library developed by Facebook that is particularly suitable for building neural networks and deep learning models. PyTorch is known for its dynamic computational graphs, which makes it very flexible in model definition and debugging.

[0028] Incremental learning: A machine learning method that allows the model to update the model without retraining the entire model when new data arrives, but instead only trains with the new data. This is useful in data streaming applications because it saves computing resources.

[0029] Streaming learning: refers to the ability of the model to continuously and in real time learn from the continuously arriving data. Similar to incremental learning, it also avoids the need to retrain the entire model, but focuses more on continuous data processing and model updates.

[0030] In order to solve the problem of poor efficiency of drug risk monitoring in the related art, the present application embodiment provides a drug risk monitoring method, which can be run on Figure 1In the computer terminal shown, the computer terminal is described below.

[0031] The drug risk monitoring method embodiment provided in the embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal for implementing a drug risk monitoring method is shown. Figure 1 As shown, the computer terminal 10 may include one or more (102a, 102b, ..., 102n are used to illustrate) processors (the processor may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission module 106 for communication functions connected via a wired and / or wireless network. In addition, it may also include: a display, a keyboard, a cursor control device, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, and a BUS bus. Those skilled in the art can understand that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0032] It should be noted that the one or more processors and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10. As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0033] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the drug risk monitoring method in the embodiment of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, realizing the above-mentioned drug risk monitoring method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0034] The transmission module 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission module 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission module 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0035] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 .

[0036] It should be noted that, in some optional embodiments, the above Figure 1 The computer terminal shown may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or a combination of hardware elements and software elements. It should be noted that Figure 1 This is merely one example of a particular embodiment and is intended to illustrate the types of components that may be present in the computer terminal described above.

[0037] Under the above-mentioned operating environment, an embodiment of the present application provides an embodiment of a drug risk monitoring method. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0038] Figure 2 is a flow chart of a drug risk monitoring method according to an embodiment of the present application, such as Figure 2As shown, the method comprises the following steps:

[0039] Step S202, obtaining individual characteristic information and initial medication data of the target subject.

[0040] In the above step S202, it is first necessary to collect individual characteristic information of the target object (usually a drug user, such as a patient), including but not limited to age, gender, disease type, genetic information, metabolic capacity, etc. At the same time, the initial medication data of the target object is obtained, that is, the drug information that the patient is currently using or about to use, such as drug name, dosage, duration of medication, etc. This information is the basis for personalized risk assessment and can reflect the patient's specific medication situation and potential risk factors. It should be noted that the source of the data involves multiple channels, including but not limited to electronic medical records, drug management records, patient self-reported information, and data extracted from non-traditional sources such as social media and online forums.

[0041] Step S204, using a risk assessment model to analyze individual characteristic information and initial medication data to obtain risk assessment results, and generate an early warning strategy based on the risk assessment results, wherein the risk assessment results are used to reflect the risk level of the initial medication data, and the risk assessment model is trained using data sources in multiple dimensions and optimized based on the feedback of the target object on the initial medication data.

[0042] In the above step S204, a risk assessment model trained by a multi-dimensional data source can be used to deeply analyze the individual characteristic information and the initial medication data. The risk assessment model is not only based on traditional medical data, but also combines non-traditional data sources such as social media, online forums, and patient evaluations, and includes a framework for a comprehensive understanding of drug risks. Through the data analysis of the risk assessment model, a risk assessment result can be obtained, which is a quantitative feedback on the degree of risk that patients may face when using specific drugs. Based on this risk assessment result, a corresponding early warning strategy can be formulated to guide patients, doctors or drug regulators to take corresponding preventive measures, such as drug-specific risk levels, possible side effects or adverse reactions, and risk triggers related to individual characteristics of patients. At the same time, the risk assessment model will also be optimized according to the patient's feedback on the initial medication data, such as adjusting its evaluation criteria and early warning generation rules according to the patient's actual response and effect, thereby improving the accuracy of the early warning.

[0043] Step S206: determining target medication data of the target subject according to the early warning strategy.

[0044] In the above step S206, after the early warning strategy is generated, the system will determine the target medication data of the target object based on the early warning strategy, including but not limited to specific measures such as adjusting the drug dosage, changing the duration of medication, recommending alternative drugs or increasing the monitoring frequency. Among them, the determination of the target medication data is a dynamic process, which will be continuously adjusted according to the patient's risk assessment results and feedback to ensure that the drugs used by the patient can both treat the disease and minimize the risk. This personalized medication guidance can significantly improve the patient's safe medication level and reduce the incidence of adverse drug events.

[0045] Through the above steps S202 to S206, the purpose of using the risk assessment model trained with multi-dimensional data sources and combining the patient's real-time feedback data to assess and warn of drug risks is achieved, thereby achieving the technical effect of refined drug risk warning and dynamic adjustment of risk strategy, thereby solving the technical problem that the drug risk monitoring methods in related technologies mostly rely on a single data source, and the risk assessment model lacks real-time and dynamic adjustment capabilities, resulting in insufficient timeliness and accuracy of risk warnings, and the inability to fully capture the real risk situation of drugs in a wide range of people and diversified usage scenarios. The following is a detailed description.

[0046] Optionally, the risk assessment model is trained in the following manner: obtaining historical drug data; extracting historical drug usage information and historical patient feedback information from the historical drug data; determining a multidimensional feature data set based on the historical drug usage information and historical patient feedback information, and dividing the multidimensional feature data set into a training set and a test set; determining an initial model, and training the initial model based on the training set, and optimizing the initial model based on the test set to obtain a risk assessment model.

[0047] In the embodiment of the present application, the training process of the risk assessment model is the core link of the entire drug risk monitoring method, which ensures that the model can make an accurate and personalized assessment of drug risks. The specific process can be as follows:

[0048] First, obtain a large amount of historical drug data, which comes from a wide range of sources, including but not limited to adverse drug reaction reports, production and sales records, clinical trial data, patient feedback, social media mentions, online forum discussions, traditional medical data, as well as scientific research papers and academic literature.

[0049] Furthermore, the quality of the historical drug data can be assessed to identify and eliminate low-quality and unreliable data, such as identifying and processing missing values, outliers, and duplicate values ​​in the data, and processing the data using data interpolation or anomaly detection technology.

[0050] Secondly, natural language processing (NLP) technology, data analysis tools and algorithms are used to extract specific information closely related to drug use and patient feedback from historical drug data, namely the above-mentioned historical drug use information and historical patient feedback information, such as drug dosage, medication duration, patient age, gender, disease type, patient satisfaction feedback on the drug, adverse reaction reports, etc.

[0051] Subsequently, the extracted historical drug use information and historical patient feedback information are integrated to form a feature data set containing multiple dimensions. The multiple dimensions can be reflected in drug characteristics (such as chemical composition, mechanism of action), individual patient characteristics (such as genetic background, physiological state) and medication context (such as dosage, duration of use).

[0052] Finally, the multi-dimensional feature dataset is divided into a training set and a test set. The training set is used to train the risk assessment model so that the model can learn how to predict drug risks based on input features, while the test set is used to evaluate the performance of the risk assessment model and verify the model's prediction accuracy on unseen data.

[0053] Specifically, before model training, you first need to set an initial model framework, which can be based on gradient boosted decision trees (GBDT) or other machine learning techniques. Subsequently, the initial model is trained using the training set. During the training process, the model will try to find the association between features and drug risks, and continuously optimize parameters to make the prediction results as close to the actual risk situation as possible. After training, the model will use the test set for performance evaluation, and adjust the model parameters and optimize according to indicators such as accuracy, recall, and F1 score to improve the prediction accuracy and stability of the model.

[0054] Through the above process, a fully trained and optimized risk assessment model can be obtained. This model can accurately assess the potential risks of drugs based on individual patient characteristics and medication use, and generate early warning strategies based on the assessment results to guide patients, doctors and regulatory agencies to take corresponding risk prevention and control measures.

[0055] Optionally, obtaining historical drug data includes: obtaining first drug data under multiple dimensions, wherein the first drug data is used to represent a single drug data corresponding to each dimension; determining a feature weight value corresponding to the first drug data, wherein the feature weight value is used to represent a contribution of the first drug data to drug risk assessment; comparing the feature weight value with a preset weight threshold, and determining the first drug data corresponding to the feature weight value exceeding the preset weight threshold as the second drug data; and fusing the second drug data to obtain historical drug data.

[0056] In the embodiments of the present application, in the process of obtaining historical drug data, a meticulous and scientific method is adopted to ensure the comprehensiveness and importance of the data, thereby effectively supporting the construction and optimization of the risk assessment model.

[0057] Specifically, first, the first drug data can be collected from multiple dimensions, which may involve drug production information, clinical application data, market feedback, patient evaluation and other aspects, ensuring multi-angle coverage of drug risk assessment. Subsequently, by determining the feature weight value corresponding to the first drug data, the contribution of each data point to drug risk assessment is quantified to help the model identify which data is the most relevant and valuable. Next, the feature weight value is compared with the preset weight threshold, and the first drug data whose weight value exceeds the threshold is screened out as the second drug data, which is considered to have a significant impact on drug risk assessment. Finally, by fusing all the screened second drug data, a historical drug data set that reflects key drug risk information can be constructed. This data set is the basis for subsequent training of the risk assessment model, ensuring that the model can learn and predict based on the most relevant and important information.

[0058] The scientificity and effectiveness of this process are reflected in the careful screening and integration of multi-source data, which avoids the negative impact of irrelevant or low-quality data on model training, while ensuring that the model can capture the most critical factors in drug risk assessment, thereby generating more accurate risk assessment results and early warning strategies. By setting feature weight values ​​and preset weight thresholds, the system can automatically identify and prioritize data that makes a significant contribution to drug risk assessment, achieving efficient training of data-driven risk assessment models, and providing a solid data foundation and technical support for real-time monitoring and early warning of drug risks.

[0059] Optionally, a multi-dimensional feature data set is determined based on historical drug usage information and historical patient feedback information, including: extracting core features from the historical drug usage information and historical patient feedback information, wherein the core features are used to represent data features related to drug risk assessment; normalizing the core features; and fusing the normalized core features to obtain a multi-dimensional feature data set.

[0060] In the embodiment of the present application, when constructing a multi-dimensional feature data set, it is also necessary to ensure the quality and applicability of the data set. The specific process can be as follows:

[0061] First, from the collected historical drug use information and historical patient feedback information, data features closely related to drug risk assessment, namely core features, are extracted through data analysis and machine learning technology. These core features may include drug dosage, duration of medication, patient age, gender, disease type, frequency of adverse reactions, patient emotional tendencies (extracted from social media and forums), etc. They can comprehensively describe multiple aspects of possible risks in the process of drug use.

[0062] Secondly, the core features are normalized to ensure that all features have the same scale during model training, avoid model learning bias caused by feature magnitude differences, and improve model stability and prediction accuracy. Specifically, normalization usually involves scaling feature values ​​to the range of 0 to 1, or standardizing them to have zero mean and unit variance, so that all features can be treated equally during model training, and the model can learn the relationship between features more effectively.

[0063] Finally, the system fuses the normalized core features to form a multi-dimensional feature data set, providing a comprehensive perspective for the model, so that the model can learn the complex associations and multi-factor influences of drug risk assessment during training. The multi-dimensional feature data set not only contains basic information on drug use, but also integrates data on individual differences in patients and market feedback, thus providing a multi-level and multi-angle data foundation for the drug risk assessment model.

[0064] Through the above process, the system can build a high-quality multi-dimensional feature data set, providing sufficient and accurate data support for subsequent risk assessment model training, ensuring that the model can predict and evaluate drug risks based on comprehensive and detailed information, and providing a solid technical foundation for intelligent monitoring and early warning of drug risks.

[0065] Optionally, the initial model is optimized based on the test set, including: using the initial model to process the test set to obtain historical risk prediction results; comparing the historical risk prediction results with actual historical risk events to obtain comparison results; when the comparison result indicates that the deviation value between the historical risk prediction results and the actual historical risk events is higher than a first preset threshold, optimizing the initial model.

[0066] In the embodiments of the present application, a feedback loop mechanism can be established, and combined with the use of test sets, to ensure objective evaluation of model performance, thereby guiding the continuous improvement of the model. Through continuous iterative optimization, the model can gradually improve its prediction accuracy of drug risks, and eventually form a mature risk assessment model that can effectively identify and assess potential drug risks, providing more accurate technical support for drug risk monitoring, early warning and prevention. The specific process can be as follows:

[0067] First, the initial model is processed using the test set to generate historical risk prediction results. The test set contains historical drug data that has not been used for model training. These data are used to test the model's predictive ability on new data. Through the model's prediction, a series of historical prediction values ​​for drug risks can be obtained.

[0068] These historical risk predictions are then compared in detail with actual historical risk events to assess the model’s prediction accuracy. Actual historical risk events refer to adverse drug reactions or other risk situations that are clearly recorded in historical data. Through this comparison, the system is able to calculate the deviation between the predictions and actual events, which is a key indicator for evaluating model performance.

[0069] If the evaluation result shows that the prediction deviation value is higher than the preset first threshold, this indicates that the model may be deficient in some aspects, such as overfitting the training data, improper feature selection, or unreasonable model parameter settings. At this time, the system will take corresponding optimization measures based on the comparison results to adjust and optimize the initial model. The optimization may include fine-tuning of model parameters, improvement of feature engineering, optimization of model algorithms, or introduction of new data sources, etc., in order to reduce prediction deviation and improve the model's generalization ability on new data.

[0070] Optionally, the above method also includes: obtaining real-time data of the target object after using the target medication data, wherein the real-time data includes real-time feedback information, real-time drug usage information and adverse reaction reports of the target object; optimizing the risk assessment model based on the real-time data, updating the data sources in multiple dimensions based on the real-time data, and improving the user interface where the target object is located based on the real-time data.

[0071] In the embodiments of the present application, a closed loop of continuous optimization is designed, and the risk assessment model and user interaction experience can be continuously improved through the collection and analysis of real-time data. Specifically, the system will collect real-time data of patients after using the target medication data regularly or in real time, such as setting up feedback forms on official websites or applications, setting up special accounts or topics, and setting up customer service hotlines or mailboxes. These data not only include real-time feedback information from patients, such as subjective evaluations such as drug efficacy and adverse reaction feelings, but also cover objective real-time drug use information and adverse reaction reports, providing direct evidence of drug risks. The acquisition of this real-time data enables the system to capture new dynamics of drugs in actual use in a timely manner, including possible new risk signals or new trends in patient groups.

[0072] Furthermore, optimizing the risk assessment model based on real-time data is a key step in the closed loop. The system inputs real-time data into the model for incremental learning or online learning, so that the model can adjust itself according to the latest drug risk information and continuously optimize the accuracy and timeliness of its risk prediction. At the same time, data sources in multiple dimensions can also be updated based on real-time data. This process ensures the real-time and comprehensiveness of the data source, provides the model with the most cutting-edge drug risk data, and further improves the accuracy of risk assessment.

[0073] In addition, improving the user interface where the target object is located based on real-time data is also an important part of user experience optimization. For example, regularly collect user feedback, including satisfaction with the query function, the convenience of using the interface, and the degree of understanding of the warning information. Based on these feedbacks, the system can improve the user interface design in a targeted manner, simplify the interaction process, and optimize the query and screening functions to ensure that patients and medical workers can obtain drug risk information and warning suggestions more efficiently and intuitively. Among them, the user interaction interface is used to provide multiple types of information and actual risk events for patients, and to display risk assessment results and warning information in the form of charts, dashboards and maps. This improvement process not only improves the user-friendliness of the system, but also enhances the communication effect of warning information, ensuring the timeliness and effectiveness of drug risk prevention and control measures.

[0074] Through continuous feedback of real-time data, the system realizes dynamic optimization of the risk assessment model and iterative improvement of the user interface, and builds an intelligent drug risk monitoring system that can respond to changes in drug risks in real time and continuously improve itself, providing strong guarantees for drug risk monitoring, early warning, and prevention and control.

[0075] Optionally, optimizing the risk assessment model based on real-time data includes: performing incremental learning on the risk assessment model using a gradient boosting decision tree based on the real-time data; determining performance indicators of the risk assessment model after performing incremental learning, wherein the performance indicators include accuracy, recall and F1 score indicators of the risk assessment model; and updating model parameters of the risk assessment model when the accuracy is lower than a second preset threshold, the recall is lower than a third preset threshold and the F1 score indicator is lower than a fourth preset threshold.

[0076] In the embodiment of the present application, in the process of optimizing the risk assessment model based on real-time data, a method combining dynamic learning and performance monitoring is adopted to ensure that the model can continuously adapt to the new data environment and the dynamic changes of drug risks. The specific process can be as follows:

[0077] First, using the data collected in real time, the gradient boosted decision tree (GBDT) is used to perform incremental learning or streaming learning on the risk assessment model. Taking incremental learning as an example, it enables the model to learn and adjust based only on the newly added data without retraining the entire data set, effectively improving the update efficiency of the model, allowing the model to quickly respond to the latest drug usage and patient feedback, and capture new risk signals in a timely manner.

[0078] After completing incremental learning, the performance indicators of the risk assessment model are evaluated, including accuracy, recall and F1 score, which are key parameters for measuring the prediction performance of the model. Accuracy reflects the proportion of events correctly predicted by the model, recall is the ability of the model to identify all actual risk events, and F1 score is the weighted harmonic average of accuracy and recall, which comprehensively reflects the prediction accuracy and comprehensiveness of the model.

[0079] If the evaluation results show that the accuracy of the model is lower than the second preset threshold, the recall rate is lower than the third preset threshold, or the F1 score is lower than the fourth preset threshold, this means that the performance of the model on real-time data has not met the expected standards, and there may be certain prediction biases or omissions. At this time, the system will update the parameters of the risk assessment model and retrain the model by adjusting the depth of the decision tree, feature selection strategy, regularization parameters, etc. to improve its prediction performance on real-time data. For example, historical data and professional opinions of domain experts set the initial thresholds and rules of the model. These thresholds and rules are determined based on the statistical analysis of past drug risk events and the expert's understanding of drug risks, providing a benchmark for the initial risk assessment of the model. When the system receives new adverse drug reaction reports, it will monitor in real time whether the number of reports exceeds the preset threshold. Once the number of reports exceeds the threshold, this is usually a sign of an increased risk signal. The system will automatically adjust the threshold of the model or trigger a higher level of warning to more sensitively capture drug risks and take timely action.

[0080] This process reflects the system's ability to self-learn and self-optimize. Through the continuous input of real-time data and regular monitoring of performance indicators, the model can continuously adjust itself and improve its ability to identify and assess drug risks. This method not only ensures the real-time and accuracy of risk assessment, but also avoids risk assessment bias caused by outdated models by dynamically adjusting model parameters, providing a more intelligent and efficient technical means for drug risk monitoring, early warning and prevention.

[0081] In the embodiments of the present application, an intelligent analysis method for drug risk monitoring, early warning and prevention is proposed, which integrates multi-source heterogeneous data (historical drug data in multiple dimensions) and uses real-time data stream analysis technology and adaptive adjustment algorithms. Through comprehensive data collection from drug production, sales, use to patient feedback, combined with emerging data sources such as social media and online forums, a more comprehensive drug risk portrait is constructed. The data is deeply analyzed using machine learning technologies such as GBDT, and the model parameters and early warning strategies are dynamically adjusted based on real-time data to ensure the real-time and accuracy of risk assessment. In addition, through the user feedback closed loop, the model and user interface are continuously optimized to improve the effectiveness of the early warning and user experience. Overall, this application provides a highly intelligent and dynamically responsive drug risk early warning system, which significantly improves the efficiency and accuracy of drug risk monitoring and provides strong technical support for drug safety supervision.

[0082] According to an embodiment of the present application, a drug risk monitoring device is provided. It should be noted that the drug risk monitoring device of the embodiment of the present application can be used to execute the drug risk monitoring method provided in the embodiment of the present application. The drug risk monitoring device provided in the embodiment of the present application is introduced below.

[0083] Figure 3 is a structural diagram of a drug risk monitoring device provided according to an embodiment of the present application. Figure 3 As shown, the device comprises:

[0084] An acquisition module 30, for acquiring individual characteristic information and initial medication data of a target subject;

[0085] The analysis module 32 is used to analyze the individual characteristic information and the initial medication data using a risk assessment model to obtain a risk assessment result, and generate an early warning strategy based on the risk assessment result, wherein the risk assessment result is used to reflect the risk level of the initial medication data. The risk assessment model is trained using data sources under multiple dimensions and is optimized based on the feedback of the target subject on the initial medication data;

[0086] The early warning module 34 is used to determine the target medication data of the target object according to the early warning strategy.

[0087] Through the acquisition module 30, analysis module 32 and warning module 34 in the above-mentioned drug risk monitoring device, the purpose of accurately predicting and dynamically preventing and controlling drug risks is achieved, thereby realizing the technical effect of improving the real-time, accuracy and personalization level of drug risk monitoring and early warning, and further solving the technical problem that the drug risk monitoring methods in related technologies mostly rely on a single data source, and the risk assessment model lacks real-time and dynamic adjustment capabilities, resulting in insufficient timeliness and accuracy of risk warnings, and unable to fully capture the real risk situation of drugs in a wide range of populations and diversified usage scenarios.

[0088] The drug risk monitoring device provided in the embodiment of the present application also includes a training module 36, which is also used to obtain historical drug data; extract historical drug usage information and historical patient feedback information from the historical drug data; determine a multi-dimensional feature data set based on the historical drug usage information and the historical patient feedback information, and divide the multi-dimensional feature data set into a training set and a test set; determine an initial model, and train the initial model based on the training set, and optimize the initial model based on the test set to obtain a risk assessment model.

[0089] In the drug risk monitoring device provided in the embodiment of the present application, the training module is also used to obtain first drug data under multiple dimensions, wherein the first drug data is used to represent a single drug data corresponding to each dimension; determine a feature weight value corresponding to the first drug data, wherein the feature weight value is used to represent the contribution of the first drug data to the drug risk assessment; compare the feature weight value with a preset weight threshold, and determine the first drug data corresponding to the feature weight value exceeding the preset weight threshold as the second drug data; and fuse the second drug data to obtain historical drug data.

[0090] In the drug risk monitoring device provided in the embodiment of the present application, the training module is also used to extract core features from historical drug usage information and historical patient feedback information, wherein the core features are used to represent data features related to drug risk assessment; the core features are normalized; and the normalized core features are fused to obtain a multi-dimensional feature data set.

[0091] In the drug risk monitoring device provided in the embodiment of the present application, the training module is also used to process the test set using the initial model to obtain historical risk prediction results; compare the historical risk prediction results with the actual historical risk events to obtain comparison results; when the comparison result indicates that the deviation value between the historical risk prediction results and the actual historical risk events is higher than a first preset threshold, optimize the initial model.

[0092] The drug risk monitoring device provided in the embodiment of the present application also includes an optimization module 38, which is also used to obtain real-time data of the target object after using the target medication data, wherein the real-time data includes real-time feedback information, real-time drug usage information and adverse reaction reports of the target object; optimize the risk assessment model based on the real-time data, update the data sources in multiple dimensions based on the real-time data, and improve the user interface where the target object is located based on the real-time data.

[0093] In the drug risk monitoring device provided in the embodiment of the present application, the optimization module is also used to perform incremental learning on the risk assessment model based on real-time data using a gradient boosting decision tree; determine the performance indicators of the risk assessment model after performing incremental learning, wherein the performance indicators include the accuracy, recall rate and F1 score indicators of the risk assessment model; and update the model parameters of the risk assessment model when the accuracy is lower than the second preset threshold, the recall rate is lower than the third preset threshold and the F1 score indicator is lower than the fourth preset threshold.

[0094] An embodiment of the present application also provides an electronic device, including: a memory and a processor, wherein the memory is used to store program instructions; the processor is connected to the memory and is used to execute the above-mentioned drug risk monitoring method.

[0095] It should be noted that the above electronic equipment is used to execute Figure 2 The drug risk monitoring method shown, therefore the relevant explanations and instructions in the above drug risk monitoring method are also applicable to the electronic device and will not be repeated here.

[0096] An embodiment of the present application also provides a non-volatile storage medium, which includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the above-mentioned drug risk monitoring method by running the computer program.

[0097] It should be noted that the above non-volatile storage medium is used to execute Figure 2 The drug risk monitoring method shown, therefore the relevant explanations and instructions in the above drug risk monitoring method are also applicable to the non-volatile storage medium and will not be repeated here.

[0098] An embodiment of the present application also provides a computer program product, including computer instructions, which implement the above-mentioned drug risk monitoring method when executed by a processor.

[0099] It should be noted that the above-mentioned computer program product is used to execute Figure 2 The drug risk monitoring method shown, therefore the relevant explanations and instructions in the above drug risk monitoring method are also applicable to the computer program product and will not be repeated here.

[0100] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0101] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0102] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0103] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0104] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0105] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0106] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A drug risk monitoring method, characterized in that: include: Obtain individual characteristic information and initial medication data of the target subject; The individual characteristic information and the initial medication data are analyzed using a risk assessment model to obtain a risk assessment result, and an early warning strategy is generated based on the risk assessment result, wherein the risk assessment result is used to reflect the risk level of the initial medication data, and the risk assessment model is trained using data sources under multiple dimensions, and is optimized based on the feedback of the target subject on the initial medication data; The target medication data of the target object is determined according to the early warning strategy.

2. The method according to claim 1, characterized in that The risk assessment model is trained in the following way: Access historical drug data; extracting historical drug use information and historical patient feedback information from the historical drug data; Determine a multi-dimensional feature data set according to the historical drug use information and the historical patient feedback information, and divide the multi-dimensional feature data set into a training set and a test set; An initial model is determined, and the initial model is trained according to the training set, and the initial model is optimized according to the test set to obtain the risk assessment model.

3. The method according to claim 2, characterized in that Access historical drug data, including: Acquire first drug data in multiple dimensions, wherein the first drug data is used to represent a single drug data corresponding to each dimension; Determining a feature weight value corresponding to the first drug data, wherein the feature weight value is used to represent a contribution of the first drug data to a drug risk assessment; Comparing the feature weight value with a preset weight threshold, and determining the first drug data corresponding to the feature weight value exceeding the preset weight threshold as the second drug data; The second drug data is integrated to obtain the historical drug data.

4. The method according to claim 2, characterized in that: Determine a multi-dimensional feature data set based on the historical drug use information and the historical patient feedback information, including: Extracting core features from the historical drug use information and the historical patient feedback information, wherein the core features are used to represent data features related to drug risk assessment; Normalizing the core features; The normalized core features are fused to obtain the multi-dimensional feature data set.

5. The method according to claim 2, characterized in that: Optimizing the initial model according to the test set includes: The test set is processed using the initial model to obtain a historical risk prediction result; Comparing the historical risk prediction results with actual historical risk events to obtain a comparison result; When the comparison result indicates that the deviation value between the historical risk prediction result and the actual historical risk event is higher than a first preset threshold, the initial model is optimized.

6. The method according to claim 1, characterized in that The method further comprises: Acquire real-time data of the target subject after using the target medication data, wherein the real-time data includes real-time feedback information, real-time medication usage information and adverse reaction reports of the target subject; The risk assessment model is optimized based on the real-time data, the data sources under the multiple dimensions are updated based on the real-time data, and the user interface where the target object is located is improved based on the real-time data.

7. The method according to claim 6, characterized in that Optimizing the risk assessment model based on the real-time data includes: Based on the real-time data, using a gradient boosting decision tree to perform incremental learning on the risk assessment model; Determining performance indicators of the risk assessment model after performing the incremental learning, wherein the performance indicators include accuracy, recall and F1 score indicators of the risk assessment model; When the accuracy is lower than a second preset threshold, the recall is lower than a third preset threshold, and the F1 score indicator is lower than a fourth preset threshold, the model parameters of the risk assessment model are updated.

8. A drug risk monitoring device, characterized in that: include: An acquisition module, used to obtain individual characteristic information and initial medication data of the target object; An analysis module, configured to analyze the individual characteristic information and the initial medication data using a risk assessment model to obtain a risk assessment result, and generate an early warning strategy based on the risk assessment result, wherein the risk assessment result is used to reflect the risk level of the initial medication data, and the risk assessment model is trained using data sources under multiple dimensions, and is optimized based on the feedback of the target subject on the initial medication data; The early warning module is used to determine the target medication data of the target object according to the early warning strategy.

9. An electronic device, characterized in that: include: A memory and a processor, wherein the memory is used to store program instructions; The processor is connected to the memory and is used to execute the drug risk monitoring method described in any one of claims 1 to 7.

10. A non-volatile storage medium, characterized in that: The non-volatile storage medium includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the drug risk monitoring method described in any one of claims 1 to 7 by running the computer program.

11. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the drug risk monitoring method described in any one of claims 1 to 7 is implemented.

Images