Active monitoring system for adverse reactions of liver cancer targeted drugs

Abstract

The application discloses a liver cancer target immune drug adverse reaction active monitoring system and belongs to the technical field of medical treatment. The liver cancer target immune drug adverse reaction active monitoring system comprises a data collection module, a data analysis module and an abnormal early warning module. The application solves the problem that the prior art cannot timely monitor patient adverse reactions. The application can construct an adverse reaction prediction model by collecting existing drug reaction data, can predict and judge real-time monitoring data monitored actively through the adverse reaction prediction model, can track the medication of patients in real time, can timely discover and handle adverse reactions, can timely send early warning information to remind medical staff to pay attention and handle when discovering that the medication of patients has abnormal conditions, and can continuously track the handling of the early warning information by the medical staff, so that the medical staff can timely handle the adverse reactions of patients.

Description

TECHNICAL FIELD

[0001] The present application relates to the medical technology field, in particular to an adverse reaction active monitoring system of liver cancer target immune drugs. BACKGROUND

[0002] Liver cancer targeted drugs are a class of drugs targeting specific targets of liver cancer, which selectively kill or inhibit the growth and division of tumor cells by acting on specific targets inside tumor cells, thereby playing a therapeutic role. The use of targeted drugs needs to be determined according to the specific situation of the patient and the characteristics of the tumor, so before use, doctors usually perform genetic testing or tumor marker testing to determine whether the patient is suitable for receiving targeted drug treatment. When using these drugs, doctors need to closely monitor the patient's condition and adverse reactions and adjust the treatment plan accordingly.

[0003] However, the existing drug adverse reaction monitoring work is mostly based on the collection of information on adverse events that have already occurred, and then unified reporting and processing after the event, which leads to a lag in the monitoring of drug adverse reactions, and the patient's adverse reactions cannot be monitored in a timely manner, so that the treatment plan cannot be adjusted in a timely and targeted manner, thus not meeting the existing needs. Therefore, we propose an adverse reaction active monitoring system of liver cancer target immune drugs. SUMMARY

[0004] The purpose of the present application is to provide an adverse reaction active monitoring system of liver cancer target immune drugs, which can construct an adverse reaction prediction model by collecting existing drug reaction data, predict and judge the real-time monitoring data actively monitored by the adverse reaction prediction model, and track the patient's medication in real time in order to timely discover and handle adverse reactions. When abnormal conditions are found in the patient's medication, an early warning information is sent in time to remind medical personnel to pay attention and handle it, and the handling of the early warning information by medical personnel is continuously tracked to ensure that medical personnel can handle the patient's adverse reactions in a timely manner, solving the problems raised in the above background technology.

[0005] To achieve the above purpose, the present application provides the following technical scheme: an adverse reaction active monitoring system of liver cancer target immune drugs, comprising:

[0006] The data collection module is used to:

[0007] Collect drug reaction data from multiple data sources, integrate and clean the collected drug reaction data, and upload them to the database;

[0008] The drug reaction data includes patient health data, target immune drug use data and adverse reaction reports.

[0009] The data analysis module is used to:

[0010] The collected drug reaction data is obtained from the database, the collected drug reaction data is processed and analyzed by using a machine learning algorithm, feature information is extracted from the drug reaction data, and patterns and features related to adverse drug reactions are identified;

[0011] An adverse reaction prediction model is established based on the analysis results of the machine learning algorithm, and the newly monitored data is predicted and judged by the adverse reaction prediction model;

[0012] An abnormality early warning module is used for:

[0013] The early warning threshold is set in advance, and based on the prediction results of the data analysis module, if the index or parameter of the prediction result exceeds the set threshold, the early warning mechanism is automatically triggered, and early warning information is sent out in time to remind medical personnel to pay attention and handle;

[0014] At the same time, the prediction results are generated into detailed reports for pharmacists and doctors to further analyze and judge.

[0015] Further, the data collection module comprises:

[0016] The data collection module is used for:

[0017] Drug reaction data is collected from various data sources using APIs, including electronic medical records, medical literature, clinical trial data, and pharmacy management systems;

[0018] The collected drug reaction data is standardized, including removing duplicates, converting data types, and handling missing values;

[0019] The integration and cleaning module is used for:

[0020] The drug reaction data is merged using ETL tools, and after integrating the data, the drug reaction data is cleaned, including checking data consistency, verifying the validity of input data, and removing erroneous records;

[0021] The data storage module is used for:

[0022] The integrated and cleaned drug reaction data is uploaded to the database for storage.

[0023] Further, the data storage module comprises:

[0024] The real-time monitoring module is used for:

[0025] The Java tool is used to monitor the data changes in the database in real time, and once the data in the database changes, the medical personnel is notified immediately;

[0026] The visualization module is used for:

[0027] The medical staff retrieves the data stored in the database according to the keywords or topics, and the medical staff is provided with various visualization display modes, including dashboards, graphs and charts, for visualizing the retrieved data.

[0028] Further, the data analysis module comprises:

[0029] The feature extraction module is configured to:

[0030] Obtain drug reaction data from the database and analyze the obtained drug reaction data using a machine learning algorithm, wherein:

[0031] Feature engineering: process the original drug reaction data to generate new features, including scaling, smoothing and binning operations;

[0032] Feature selection: select a feature subset according to the performance indicators of the constructed prediction model, and the feature subset is selected by using the search and sorting of the SelectK library;

[0033] The model construction module is configured to:

[0034] Combine the selected machine learning algorithm and the feature selection result of the feature extraction module to construct an adverse reaction prediction model;

[0035] The training optimization module is configured to:

[0036] Train the constructed adverse reaction prediction model using the training set and verify it on the test set;

[0037] If the adverse reaction prediction model performs poorly, the adverse reaction prediction model is fine-tuned through grid search, random search and Bayesian optimization methods;

[0038] The model application module is configured to:

[0039] Deploy the trained adverse reaction prediction model to the actual environment;

[0040] Active monitoring of patients to obtain real-time monitoring data of patients, and predicting and judging the active monitoring real-time monitoring data through the adverse reaction prediction model.

[0041] Further, the abnormal early warning module comprises:

[0042] The early warning notification module is configured to:

[0043] Set the early warning threshold in advance and based on the prediction result of the data analysis module;

[0044] If the index or parameter of the prediction result exceeds the set threshold, the early warning mechanism is automatically triggered, and early warning information is sent to medical personnel through multiple channels;

[0045] The early warning tracking module is configured to:

[0046] The processing of the early warning information by the medical personnel is continuously tracked, and the early warning mechanism is adjusted and optimized according to the feedback;

[0047] The report generation module is configured to:

[0048] The prediction result is generated into a detailed report including the early warning reason and the prediction result, for further analysis and judgment by pharmacists and doctors.

[0049] Further, the early warning notification module comprises:

[0050] The early warning setting module is configured to:

[0051] The early warning threshold is set in advance, including the type of threshold, the calculation method of threshold and the way of early warning;

[0052] The result acquisition module is configured to:

[0053] After the data analysis module makes a prediction, the prediction result of the data analysis module is acquired, and the prediction result is compared with the pre-set threshold to determine whether the early warning needs to be triggered;

[0054] The information sending module is configured to:

[0055] When the index or parameter of the prediction result exceeds the set threshold, the early warning mechanism is automatically triggered and early warning information is generated, including the early warning level, the early warning content and the early warning time, and the early warning information is pushed through the pre-set channel.

[0056] Further, the early warning tracking module comprises:

[0057] The early warning monitoring module is configured to:

[0058] The processing progress of the early warning information by the medical personnel is monitored, including the number of early warnings that have been processed, the number of early warnings that have been closed and the number of early warnings that have not been processed, and the processing status of each early warning;

[0059] The feedback collection module is configured to:

[0060] The feedback information of the medical personnel on the early warning information is collected, including the specific content of the feedback, the time of the feedback and the identity information of the feedback person;

[0061] The mechanism optimization module is configured to:

[0062] According to the collected feedback information and monitored data, the early warning mechanism is adjusted and optimized, including the adjustment of threshold and the modification of early warning channel.

[0063] Further, the report generation module has the following functions:

[0064] Report conversion function: convert the generated report into specific report document formats, including PDF, Word and Excel;

[0065] Add annotation function: add annotations to the generated report, including explanations and analysis of the prediction results, and suggestions for subsequent actions;

[0066] Export and sharing function: provide report export and sharing function for pharmacists and doctors to export the report to local or other systems for viewing, or share the report with others for discussion and decision-making.

[0067] Further, the running state of the report sharing is monitored in real time, and when the sharing operation is abnormal, an abnormal alarm is given, including:

[0068] Collect the number of data sharing times in each monitoring period;

[0069] Compare the number of data sharing times with the preset sharing times threshold;

[0070] When the number of data sharing times exceeds the preset sharing times threshold, the actual report sharing display data volume, report garbled rate and the number of unit times experienced by the data sharing completion in each unit time during each report sharing process are retrieved;

[0071] Extract the data volume completed by sharing in each unit time during each data sharing process;

[0072] Obtain the first report sharing running coefficient corresponding to each data sharing using the data volume completed by sharing in each unit time;

[0073] Wherein, the first report sharing running coefficient is obtained by the following formula:

[0074]

[0075] Wherein, B 01 represents the first report sharing running coefficient; n represents the number of unit times experienced by each data sharing; D i represents the data volume completed by sharing corresponding to the ith unit time; T d represents the length of unit time; D p represents the average value of the data volume completed by sharing corresponding to n unit times; D drepresents the standard deviation of the amount of data completed sharing corresponding to n unit time; D max represents the maximum value of the amount of data completed sharing corresponding to n unit time;

[0076] The first report sharing running coefficient is used in combination with the actual report sharing display data amount, the report error code rate and the number of unit times experienced by the data sharing completion corresponding to each unit time of sharing to make an abnormality determination on the running state of the report sharing.

[0077] Further, the first report sharing running coefficient is used in combination with the actual report sharing display data amount, the report error code rate and the number of unit times experienced by the data sharing completion corresponding to each unit time of sharing to make an abnormality determination on the running state of the report sharing, comprising:

[0078] The actual report sharing display data amount, the report error code rate and the number of unit times experienced by the data sharing completion corresponding to each unit time of sharing are used in combination with the first report sharing running coefficient to obtain a second report sharing running coefficient;

[0079] The second report sharing running coefficient is obtained through the following formula:

[0080]

[0081] The second report sharing running coefficient is obtained through the following formula: 02 represents the second report sharing running coefficient; m represents the number of data sharing times; P wj represents the report error code rate corresponding to the jth data sharing; B 01j represents the first report sharing running coefficient corresponding to the jth data sharing; n represents the number of unit times experienced by each data sharing; D i represents the amount of data completed sharing corresponding to the ith unit time; D z represents the overall data amount corresponding to the jth data sharing; P si represents the actual report sharing display data amount corresponding to the amount of data completed sharing of the ith unit time contained in the jth data sharing; P xi represents the theoretical report sharing display data amount corresponding to the amount of data completed sharing of the ith unit time contained in the jth data sharing;

[0082] The second report sharing running coefficient is compared with a preset sharing running coefficient threshold value;

[0083] When the second report sharing running coefficient is lower than the preset sharing running coefficient threshold value, it is determined that the data sharing running is abnormal.

[0084] Compared with the prior art, the present application has the following beneficial effects:

[0085] 1、The present application can construct an adverse reaction prediction model by collecting existing drug reaction data, and simultaneously actively monitor patients to obtain real-time monitoring data of the patients, predict and judge the real-time monitoring data actively monitored by the adverse reaction prediction model, and track the medication of the patients in real time to discover and handle adverse reactions in a timely manner.

[0086] 2、The present application compares and analyzes the prediction results, and if the indexes or parameters of the prediction results are abnormal, an early warning mechanism is automatically triggered and early warning information is sent to medical personnel to remind the medical personnel to pay attention and handle, and the handling of the early warning information by the medical personnel is continuously tracked to ensure that the medical personnel timely handle the adverse reactions of the patients. BRIEF DESCRIPTION OF DRAWINGS

[0087] Figure 1 The present application is a structure diagram of an active monitoring system for adverse reactions of liver cancer target immune drugs. DETAILED DESCRIPTION

[0088] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to 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. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0089] To solve the technical problem that the existing adverse reaction monitoring of drugs is mostly based on the information collection of adverse events of patients, and then the unified reporting and processing are performed after the event, which leads to the lag of the monitoring of adverse reactions of drugs and the inability to timely monitor the adverse reactions of patients, so that the treatment plan cannot be adjusted in a targeted manner in a timely manner, please refer to Figure 1 The present application provides the following technical solutions:

[0090] The active monitoring system for adverse reactions of liver cancer target immune drugs comprises:

[0091] The data collection module is used for:

[0092] Collecting drug reaction data from multiple data sources, integrating and cleaning the collected drug reaction data, and uploading the drug reaction data to a database;

[0093] The drug reaction data comprises patient health data, target immune drug use data, and adverse reaction reports.

[0094] The data analysis module is used for:

[0095] Obtaining collected drug reaction data from a database and processing and analyzing it using machine learning algorithms (such as deep learning, random forests, support vector machines, etc.). These algorithms can extract useful information from large amounts of drug reaction data and identify patterns and features related to adverse drug reactions;

[0096] An adverse reaction prediction model is established based on the analysis results of the machine learning algorithm. The model is used to predict and judge the newly monitored data, thereby achieving automatic early warning of adverse reactions to liver cancer target immune drugs.

[0097] Abnormal warning module, used for:

[0098] Set warning thresholds in advance. Based on the prediction results of the data analysis module, if the indicators or parameters of the prediction results exceed the set thresholds, the warning mechanism will be automatically triggered and warning information will be issued in time to remind medical staff to pay attention and take action;

[0099] At the same time, a detailed report will be generated based on the prediction results for further analysis and judgment by pharmacists and doctors.

[0100] The technical effects of the above content are as follows: the data collection module is responsible for collecting drug reaction data from multiple data sources, integrating and cleaning it, and uploading it to the database; the data analysis module uses machine learning algorithms, such as deep learning, random forests, support vector machines, etc., to analyze the collected drug reaction data to extract useful information and identify patterns and characteristics related to adverse drug reactions. Based on the analysis results of the above algorithms, an adverse reaction prediction model can be established to predict and judge the newly monitored data, thereby realizing automatic early warning of adverse reactions of liver cancer target-immune drugs; the abnormal early warning module will automatically trigger the early warning mechanism based on the prediction results of the data analysis module when the indicators or parameters of the prediction results exceed the set threshold, and issue early warning information in time to remind medical staff to pay attention and deal with it. In addition, the early warning results will also generate a detailed report for pharmacists and doctors to conduct further analysis and judgment. Through the above operations, through active monitoring of patients, the early warning and prevention capabilities of adverse drug reactions can be effectively improved, providing important guarantees for clinical drug safety.

[0101] Data collection module, including:

[0102] Data acquisition module for:

[0103] Use APIs to collect drug response data from various data sources, including electronic medical records, medical literature, clinical trial data, and pharmacy management systems;

[0104] standardizing the collected drug reaction data to ensure comparability, including removing duplicates, converting data types (e.g., converting all text descriptions to coded values), and handling missing values;

[0105] the integration cleaning module is used to:

[0106] using ETL (Extract, Transform, Load) tools to merge drug reaction data, after integrating the data, the drug reaction data needs to be cleaned to eliminate any errors or irrelevant records, including checking data consistency, verifying the validity of input data, and removing erroneous records;

[0107] the data storage module is used to:

[0108] uploading the integrated and cleaned drug reaction data to the database for storage.

[0109] The technical effects of the above are: the data collection module uses API interfaces to extract the required drug reaction data from various data sources, and secondly, to ensure the accuracy and comparability of the data, the collected drug reaction data needs to be standardized, such as removing duplicates, converting data types (e.g., converting text descriptions to coded values), and handling missing values. The integration cleaning module uses ETL (Extract, Transform, Load) tools to merge and clean the standardized data, after merging the data, the consistency of the data needs to be carefully checked to ensure the correctness of the data, at the same time, the validity of the input data is verified to avoid the interference of invalid data, in this process, if there is erroneous data, it will be removed in time to ensure that the integration cleaning module is pure and reliable, and the data storage module uploads the cleaned and arranged drug reaction data to the database for storage, providing a basis for our subsequent analysis.

[0110] the data storage module includes:

[0111] the real-time monitoring module is used to:

[0112] using Java tools to monitor the data changes in the database in real time, and once the data in the database changes, the medical staff will be notified immediately;

[0113] the visualization module is used to:

[0114] medical staff can search the stored data in the database according to keywords or topics, and provide multiple visualization display methods for medical staff, including dashboards, graphics, and charts, to visualize the retrieved data;

[0115] In addition, the data storage module also has the following functions:

[0116] Backup and recovery function: regular backup of drug reaction data in the database, such as daily, weekly or monthly, when data loss or damage occurs, the lost or damaged data can be recovered from the backup;

[0117] Secure access function: identity verification and authorization for medical personnel accessing the database, only authorized medical personnel can access and operate the database.

[0118] The technical effects of the above are: the real-time monitoring module can be used to monitor the data changes in the database in real time, which is implemented by Java tools, and can immediately notify the medical personnel once the database data changes, so that the medical personnel can timely understand the latest data changes, and each new data addition will increase the sample size of the model, thereby improving the generalization ability and accuracy of the model, when there are more data, the model can better capture the potential patterns and rules of the data, in order to continuously improve the prediction performance of the model, the visualization module allows medical personnel to search by keywords or topics, so as to facilitate finding the specific data needed, and provides a variety of visualization display methods, so that medical personnel can more intuitively view the searched data, further, the backup and recovery function ensures the security of drug reaction data, by regularly backing up the data in the database, such as daily, weekly or monthly, to prevent data loss or damage, and once data anomaly occurs, the data can be recovered from the backup, improving the security of the data, the secure access function ensures that only authorized medical personnel can access and operate the database, effectively preventing unauthorized access, and protecting sensitive information from being leaked.

[0119] Data analysis module, including:

[0120] Feature extraction module, used for:

[0121] Obtain drug reaction data from the database and use machine learning algorithms to analyze the obtained drug reaction data to extract useful information and identify relevant patterns and features, wherein:

[0122] Feature engineering: process the original drug reaction data to generate new features, including scaling, smoothing and binning operations, which can better reveal the internal rules of the data and provide more effective features for subsequent modeling;

[0123] Feature selection: select a subset of features according to the performance indicators (such as precision, recall, F1 score, etc.) of the constructed prediction model, which can use SelectK library for fast search and sorting to reduce the overhead of computation time and space;

[0124] Model construction module, used for:

[0125] Combine the feature selection results of the selected machine learning algorithm and the feature extraction module to build an adverse reaction prediction model;

[0126] The training optimization module is used for:

[0127] The training set is used to train the constructed adverse reaction prediction model, and the test set is used for verification;

[0128] If the adverse reaction prediction model performs poorly, methods such as grid search, random search, and Bayesian optimization can be used to optimize the adverse reaction prediction model to improve prediction accuracy;

[0129] The model application module is used for:

[0130] The trained adverse reaction prediction model is deployed to the actual environment;

[0131] Active monitoring of patients is carried out to obtain real-time monitoring data of patients, and the real-time monitoring data actively monitored by the adverse reaction prediction model is predicted and judged.

[0132] The technical effects of the above are: The feature extraction module processes the original drug reaction data to generate new features, such as scaling, smoothing, and binning, to enable machine learning algorithms to better learn and understand these data. The model construction module combines the selected machine learning algorithm and the feature selection results of the feature extraction module to build an adverse reaction prediction model. The training optimization module uses the training set to train the constructed adverse reaction prediction model and verifies it on the test set. If the model performs poorly, it can be optimized through methods such as grid search, random search, and Bayesian optimization to improve prediction accuracy. Finally, the model application module deploys the trained adverse reaction prediction model to the actual environment and actively monitors patients to obtain real-time monitoring data of patients. The real-time monitoring data actively monitored by the adverse reaction prediction model is predicted and judged. Based on the above operations, active monitoring of patients is carried out, so that the constructed adverse reaction prediction model can be used to predict the data actively monitored, potential adverse drug reactions can be discovered and measures can be taken in time to avoid potential adverse drug reactions, further protecting the health of patients, and doctors can also more accurately predict adverse drug reactions, thereby reducing the risk of medication errors.

[0133] The abnormality warning module includes:

[0134] The warning notification module is used for:

[0135] The warning threshold is set in advance, and the prediction result of the data analysis module is based on;

[0136] If the indicators or parameters of the predicted results exceed the set threshold, the early warning mechanism will be automatically triggered and warning information will be sent to medical staff through multiple channels to remind them to pay attention and take appropriate measures;

[0137] Early warning tracking module, used for:

[0138] Continuously track how medical staff handle early warning information, and adjust and optimize the early warning mechanism based on feedback;

[0139] Report generation module for:

[0140] The prediction results will be generated into a detailed report, including the warning reasons and prediction results, for pharmacists and doctors to conduct further analysis and judgment.

[0141] The technical effects of the above content are as follows: the early warning notification module is responsible for setting the early warning threshold in advance, and based on the prediction results of the data analysis module, when the indicators or parameters of the prediction results exceed the set threshold, the early warning mechanism is automatically triggered and early warning information is sent to medical staff through multiple channels to remind them to pay attention and handle the situation. The early warning tracking module is responsible for continuously tracking the medical staff's handling of the early warning information and adjusting and optimizing the early warning mechanism based on the feedback to ensure that medical staff respond to the issued early warning information in a timely manner. The report generation module is responsible for converting the prediction results into detailed reports so that pharmacists and doctors can conduct in-depth analysis and judgment. Based on the above design, the setting of the early warning threshold can be dynamically adjusted according to historical data and current conditions to adapt to different environmental changes. The early warning information can be released through multiple channels, ensuring that medical staff can receive the early warning information in a timely manner regardless of where they are. By continuously tracking and providing feedback on the medical staff's handling of the early warning information, the effectiveness of the early warning mechanism can be understood in real time and corresponding adjustments and optimizations can be made. These key points can help decision makers better understand and evaluate the value and effectiveness of the abnormal early warning module. The abnormal early warning module can effectively improve the quality and efficiency of medical services and reduce error rates and risks.

[0142] Early warning notification module, including:

[0143] The early warning setting module is used to:

[0144] Set warning thresholds in advance, including the type of threshold (such as univariate thresholds and multivariate thresholds), the calculation method of the threshold (such as mean, median, and mode), and the warning method (such as email notification and SMS push);

[0145] The result acquisition module is used to:

[0146] After the data analysis module makes a prediction, the prediction result of the data analysis module is obtained and compared with the pre-set threshold to determine whether an early warning needs to be triggered;

[0147] an information sending module, configured to:

[0148] When the index or parameter of the prediction result exceeds the set threshold, an early warning mechanism is automatically triggered and early warning information is generated, including early warning level, early warning content, early warning time and other information, which is pushed through a pre-set channel (such as email and short message, etc.).

[0149] The technical effects of the above content are: the main task of the early warning setting module is to set the early warning threshold and the early warning method, which can meet the needs of users in different situations, the main task of the result obtaining module is to obtain the prediction result after the data analysis module makes a prediction, and then compare the prediction result with the pre-set threshold, if the prediction result exceeds the threshold, it means that there may be an abnormal situation, at this time, the early warning mechanism needs to be triggered, the main task of the information sending module is to automatically trigger the early warning mechanism and generate early warning information when the index or parameter of the prediction result exceeds the set threshold, the content of the early warning information should include early warning level, early warning content, early warning time and other information, which can enable medical staff to better understand and respond to early warning situations, and the early warning information will be pushed through a pre-set channel, such as email and short message, etc., which can ensure the timeliness and effectiveness of the early warning information.

[0150] an early warning tracking module, configured to:

[0151] an early warning monitoring module, configured to:

[0152] monitor the processing progress of medical staff on early warning information, including the number of early warnings that have been processed, the number of early warnings that have been closed, the number of early warnings that have not been processed, and the processing status of each early warning (such as unprocessed, processed, pending review, etc.);

[0153] a feedback collection module, configured to:

[0154] collect feedback information of medical staff on early warning information, including the specific content of feedback, the time of feedback, the identity of the feedback person and other information, the feedback information can understand the actual effect and use of early warning, and then optimize the early warning mechanism;

[0155] a mechanism optimization module, configured to:

[0156] According to the collected feedback information and monitored data, the early warning mechanism is appropriately adjusted and optimized to improve the accuracy and timeliness of early warning, the adjustment and optimization may include adjustment of threshold and modification of early warning channel (and modification of early warning rules and improvement of early warning mechanism).

[0157] The technical effects of the above are: the early warning monitoring module is mainly responsible for monitoring the processing progress of medical staff on early warning information, including but not limited to the number of early warnings that have been processed, the number of early warnings that have been closed, the number of early warnings that have not been processed, and the processing status of each early warning (such as unprocessed, processed, and pending review), so as to timely grasp the actual processing situation of early warning information, thereby adjusting the early warning mechanism, the feedback collection module is responsible for collecting the feedback information of medical staff on early warning information, the feedback information can deeply understand the actual effect and use of early warning, thereby optimizing and improving the early warning mechanism, the mechanism optimization module is responsible for adjusting and optimizing the early warning mechanism according to the collected feedback information and monitored data, involving the adjustment of threshold and the modification of early warning channel, and even including the modification of early warning rules and the improvement of early warning mechanism, through adjustment and optimization, aiming to improve the accuracy and timeliness of early warning, thereby better serving medical workers.

[0158] The report generation module has the following functions:

[0159] Report conversion function: converting the generated report into specific report documents, including PDF, Word, Excel and other formats, which can be easily printed;

[0160] Add annotation function: add annotations in the generated report to guide pharmacists and doctors to further analyze and judge the prediction results, the annotations include the explanation and analysis of the prediction results, and the suggestions for the subsequent actions;

[0161] Export and sharing function: provide the export and sharing function of the report, so that pharmacists and doctors can export the report to the local or other system for viewing, and also can share the report with others for discussion and decision-making.

[0162] The technical effects of the above are: the report conversion function is mainly to convert the generated report into multiple forms such as PDF, Word, Excel, etc., which not only facilitates medical staff to print directly, but also makes the report to be saved and used in a more rich form, the purpose of adding annotation function is to help pharmacists and doctors better understand the content of the report and make more in-depth analysis on the prediction results, adding annotations in the report not only includes the explanation and analysis of the prediction results, but also includes the suggestions for the subsequent actions, which can make the use of the report more clear and efficient, the export and sharing function is to let medical staff can easily export the report to the local or other system for viewing, and also can share the report with others for discussion and decision-making, making the use scenario of the report more extensive, not limited to single medical staff, and also makes the use process of the report more flexible.

[0163] Working principle: through the data collection module from multiple data sources collect drug reaction data and upload to the database, data analysis module uses machine learning algorithm to analyze the collected drug reaction data, based on the analysis results of the above algorithm, an adverse reaction prediction model can be established to predict and judge new monitored data, realize the automatic warning of hepatocellular carcinoma target drug adverse reactions, the abnormal warning module can automatically trigger the warning mechanism according to the prediction results of the data analysis module, and timely send warning information to remind medical personnel to pay attention and handle, in addition, through the continuous tracking and feedback of the medical personnel's handling of the warning information, the effect of the warning mechanism can be understood in real time, through the above operation, through the active monitoring of patients, the warning and prevention ability of drug adverse reactions can be effectively improved, and important guarantee for clinical drug safety is provided.

[0164] Specifically, the running state of real-time monitoring report sharing is monitored, and abnormal alarm is performed when the shared running is abnormal, including:

[0165] Real-time acquisition of data sharing times in each monitoring period;

[0166] Comparing the data sharing times with the preset sharing times threshold;

[0167] When the data sharing times exceed the preset sharing times threshold, the actual report sharing display data volume, report garbled code rate and the number of unit times experienced by data sharing in each unit time in each report sharing process are called;

[0168] Extracting the data volume shared in each unit time in each data sharing process;

[0169] Obtaining the first report sharing running coefficient corresponding to each data sharing by using the data volume shared in each unit time;

[0170] Wherein, the first report sharing running coefficient is obtained by the following formula:

[0171]

[0172] Wherein, B 01 represents the first report sharing running coefficient; n represents the number of unit times experienced by each data sharing; D i represents the data volume corresponding to the i th unit time completed sharing; T d represents the length of unit time; D p represents the average value of the data volume corresponding to n unit times completed sharing; D d represents the standard deviation of the data volume corresponding to n unit times completed sharing; D max represents the maximum value of the data volume corresponding to n unit times completed sharing;

[0173] The first report sharing running coefficient is used in combination with the actual report sharing display data amount shared per unit time, the report garbled code rate, and the number of unit times experienced by data sharing completion to perform abnormality determination on the running state of report sharing.

[0174] The technical effects of the above technical solution are as follows: by collecting the number of data sharing times in each monitoring period in real time, the system can continuously monitor the running state of report sharing, ensuring comprehensive understanding of the sharing activity. When the number of data sharing times exceeds the preset sharing time threshold, the system can quickly respond and further analyze specific indicators in the sharing process, thereby realizing timely discovery and processing of abnormal conditions. The system not only focuses on the number of sharing times, but also deeply analyzes multiple key indicators in each report sharing process, including the actual report sharing display data amount shared per unit time, the report garbled code rate, and the number of unit times experienced by data sharing completion. By introducing the first report sharing running coefficient (B01), multiple dimensions such as the amount of data completed per unit time, the standard deviation of the data amount, the maximum value, and the average value are comprehensively considered, making the abnormality determination more accurate and comprehensive. Extracting the amount of data completed per unit time in each data sharing process and using these data for subsequent analysis and calculation demonstrates the system's powerful data processing capabilities. The formulaic calculation method makes the abnormality determination process more objective and scientific, reducing the subjectivity and uncertainty of human judgment. Through real-time monitoring and abnormality determination, the system can timely discover and process abnormal conditions in the sharing process, thereby avoiding the impact of potential faults and problems on system stability and reliability. This helps to improve the overall performance and user experience of the system, ensuring the smooth progress of report sharing activities. Through real-time monitoring and abnormality determination of the running state of report sharing, the system can provide strong support for resource allocation and decision-making.

[0175] In summary, the technical solution realizes effective monitoring and management of the running state of report sharing through real-time monitoring, accurate abnormality determination, efficient data processing, improvement of system stability and reliability, and optimization of resource allocation and decision support.

[0176] Specifically, the first report sharing running coefficient is used in combination with the actual report sharing display data amount shared per unit time, the report garbled code rate, and the number of unit times experienced by data sharing completion to perform abnormality determination on the running state of report sharing, including:

[0177] The actual report sharing display data amount shared per unit time, the report garbled code rate, and the number of unit times experienced by data sharing completion are used in combination with the first report sharing running coefficient to obtain a second report sharing running coefficient;

[0178] The second report sharing operation coefficient is obtained by the following formula:

[0179]

[0180] B 02 represents the second report sharing operation coefficient; m represents the number of data sharing times; P wj represents the report error code rate corresponding to the jth data sharing; B 01j represents the first report sharing operation coefficient corresponding to the jth data sharing; n represents the number of unit times experienced per data sharing; D i represents the data amount completed sharing in the ith unit time; D z represents the overall data amount corresponding to the jth data sharing; P si represents the actual report sharing display data amount corresponding to the data amount completed data sharing in the ith unit time contained in the jth data sharing; P xi represents the theoretical report sharing display data amount corresponding to the data amount completed data sharing in the ith unit time contained in the jth data sharing;

[0181] The second report sharing operation coefficient is compared with a preset sharing operation coefficient threshold value;

[0182] When the second report sharing operation coefficient is lower than the preset sharing operation coefficient threshold value, it is determined that data sharing operation is abnormal.

[0183] The technical effects of the above technical solution are: the scheme introduces a second report sharing operation coefficient (B02), comprehensively considers the actual report sharing display data volume corresponding to each unit time, the report garbled code rate, the number of unit times experienced during data sharing, and the first report sharing operation coefficient, and the like, so that the running state of the report sharing can be more comprehensively and accurately evaluated. The second report sharing operation coefficient is calculated using a complex formula and multiple parameters, making the abnormality determination more accurate. This helps to reduce false positives and false negatives, and improves the reliability and stability of the system. The scheme can sensitively capture abnormal changes in data volume, garbled code rate, and data sharing time during the report sharing process, thereby issuing an abnormal alarm in a timely manner to ensure that administrators can respond quickly and handle problems. Through real-time monitoring and abnormality determination, the system can provide administrators with comprehensive information about the running state of the report sharing, which helps administrators optimize resource allocation and make more informed decisions. For example, when an abnormal running state of a data sharing is found, the administrator can adjust the sharing strategy or increase resource investment in a timely manner to ensure the smooth progress of the sharing activity. Timely detection and handling of abnormal situations during the report sharing process helps to improve user experience and system performance. This can reduce inconvenience and losses caused by sharing failures or data errors, and improve user satisfaction and trust in the system. The scheme evaluates the running state of the report sharing by introducing multiple parameters and complex calculation formulas, which makes the system more adaptive and scalable. As technology evolves and user needs change, administrators can adjust parameters and formulas according to actual conditions to adapt to new sharing scenarios and needs.

[0184] In summary, the technical solution comprehensively evaluates the running state of the report sharing, improves the accuracy of abnormality determination, sensitively captures abnormal changes, optimizes resource allocation and decision support, improves user experience and system performance, and enhances the adaptability and scalability of the system, thereby achieving effective monitoring and management of the running state of the report sharing.

[0185] It should be noted that in this paper, relationship terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed or inherent to such a process, method, article or device.

[0186] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and changes can be made by those skilled in the art without departing from the spirit and principles of the present application.

Claims

1. Active monitoring system for adverse reactions of liver cancer target immune drugs, characterized by: include: Data collection module for: Collect drug response data from multiple data sources, integrate and clean the collected drug response data and upload them to the database; Among them, drug response data include patient health data, target immune drug usage data and adverse reaction reports; Data analysis module for: Obtaining collected drug reaction data from a database, processing and analyzing the collected drug reaction data using machine learning algorithms, extracting feature information from the drug reaction data, and identifying patterns and features related to adverse drug reactions; Establish an adverse reaction prediction model based on the analysis results of the machine learning algorithm, and use the adverse reaction prediction model to predict and judge the newly monitored data; Abnormal warning module, used for: Set warning thresholds in advance. Based on the prediction results of the data analysis module, if the indicators or parameters of the prediction results exceed the set thresholds, the warning mechanism will be automatically triggered and warning information will be issued in time to remind medical staff to pay attention and take action; At the same time, the prediction results will be generated into a detailed report, which can be exported and shared for further analysis and judgment by pharmacists and doctors. Among them, real-time monitoring and reporting of the shared operation status, and abnormal alarm when the shared operation is abnormal, including: Real-time collection of data sharing times within each monitoring cycle; Comparing the data sharing times with a preset sharing times threshold; When the data sharing times exceed the preset sharing times threshold, the actual report sharing display data volume, report garbled code rate and the number of unit time experienced by data sharing completion corresponding to each unit time sharing in each report sharing process are retrieved; Extract the amount of data shared per unit time during each data sharing process; The first report sharing operation coefficient corresponding to each data sharing is obtained by using the data amount shared per unit time; The first report shared operation coefficient is obtained by the following formula: Among them, B 01 represents the first report sharing operation coefficient; n represents the number of unit times experienced by each data sharing; D i T represents the amount of data shared per unit time; d Indicates the length of unit time; D p represents the average amount of shared data corresponding to n unit time; D d represents the standard deviation of the amount of shared data corresponding to n unit time; D max Indicates the maximum amount of shared data that occurs in n units of time; Determine the abnormality of the report sharing operation state by using the first report sharing operation coefficient in combination with the actual report sharing display data volume corresponding to each unit time sharing, the report garbled code rate, and the number of unit times experienced by data sharing completion; The abnormality determination of the operating status of the report sharing is performed using the first report sharing operating coefficient in combination with the actual report sharing display data volume corresponding to each unit time sharing, the report garbled code rate, and the number of unit times experienced by the data sharing completion, including: Obtaining a second report sharing operation coefficient by combining the actual report sharing display data volume corresponding to each unit time sharing, the report garbled code rate, and the number of unit times experienced for completing data sharing with the first report sharing operation coefficient; The second report shared operation coefficient is obtained by the following formula: Among them, B 02 represents the second report sharing operation coefficient; m represents the number of data sharing times; P wj represents the reported garbled code rate corresponding to the jth data sharing; B 01j represents the first report sharing operation coefficient corresponding to the jth data sharing; n represents the number of unit time experienced by each data sharing; D i Indicates the amount of data shared per unit time; D z represents the total data volume corresponding to the jth data sharing; P si P represents the actual report sharing display data volume corresponding to the data volume completed in the i-th unit time contained in the j-th data sharing; xi The theoretical report sharing display data volume corresponding to the data volume completed in the i-th unit time contained in the j-th data sharing; comparing the second reported shared operating coefficient with a preset shared operating coefficient threshold; When the second reported sharing operation coefficient is lower than a preset sharing operation coefficient threshold, it is determined that the data sharing operation is abnormal.

2. The liver cancer target immune drug adverse reaction active monitoring system according to claim 1, characterized in that: The data collection module includes: Data acquisition module for: Use APIs to collect drug response data from various data sources, including electronic medical records, medical literature, clinical trial data, and pharmacy management systems; Standardize the collected drug response data, including removing duplicates, converting data types, and handling missing values; Integrated cleaning module for: Use ETL tools to merge drug response data. After data integration, clean the drug response data, including checking data consistency, verifying the validity of input data, and removing erroneous records. Data storage module for: The integrated and cleaned drug response data are uploaded to the database for storage.

3. The liver cancer target immune drug adverse reaction active monitoring system according to claim 2, characterized in that: The data storage module includes: Real-time monitoring module for: Use Java tools to monitor data changes in the database in real time, and immediately notify medical staff if any changes are found in the database; Visualization module for: Medical staff can retrieve data stored in the database based on keywords or topics, and a variety of visual display methods are provided to medical staff, including dashboards, graphs and charts, to visualize the retrieved data.

4. The liver cancer target immune drug adverse reaction active monitoring system according to claim 1, characterized in that: The data analysis module includes: Feature extraction module, used to: Obtain drug response data from a database and analyze the obtained drug response data using a machine learning algorithm, wherein: Feature engineering: Processing the raw drug response data to generate new features, including scaling, smoothing, and binning operations; Feature selection: Select a feature subset based on the performance indicators of the constructed prediction model. The feature subset is selected by searching and sorting using the SelectK library; Model building modules for: Combining the selected machine learning algorithm with the feature selection results of the feature extraction module to build an adverse reaction prediction model; Training optimization module, used to: The constructed adverse reaction prediction model was trained using the training set and verified on the test set; If the adverse reaction prediction model performs poorly, the adverse reaction prediction model will be optimized through grid search, random search and Bayesian optimization methods; Model application module, used for: Deploy the trained adverse reaction prediction model into the actual environment; Actively monitor patients to obtain real-time monitoring data of patients, and use adverse reaction prediction models to predict and judge the real-time monitoring data obtained through active monitoring.

5. The liver cancer target immune drug adverse reaction active monitoring system according to claim 1, characterized in that: The abnormal warning module includes: Early warning notification module, used to: Set warning thresholds in advance and based on the prediction results of the data analysis module; If the indicators or parameters of the predicted results exceed the set threshold, the early warning mechanism will be automatically triggered and early warning information will be sent to medical staff through multiple channels; Early warning tracking module, used for: Continuously track how medical staff handle early warning information, and adjust and optimize the early warning mechanism based on feedback; Report generation module for: The prediction results will be generated into a detailed report, including the warning reasons and prediction results, for pharmacists and doctors to conduct further analysis and judgment.

6. The liver cancer target immune drug adverse reaction active monitoring system according to claim 5, characterized in that: The early warning notification module includes: The early warning setting module is used to: Set warning thresholds in advance, including the type of threshold, how to calculate the threshold, and how to issue warnings; The result acquisition module is used to: After the data analysis module makes a prediction, the prediction result of the data analysis module is obtained and compared with the pre-set threshold to determine whether an early warning needs to be triggered; Information sending module, used for: When the indicators or parameters of the prediction results exceed the set threshold, the early warning mechanism is automatically triggered and early warning information is generated, including the warning level, warning content and warning time, and the early warning information is pushed through pre-set channels.

7. The liver cancer target immune drug adverse reaction active monitoring system according to claim 5, characterized in that: The early warning tracking module includes: Early warning monitoring module, used for: Monitor the progress of medical staff in handling warning information, including the number of warnings that have been handled, the number of warnings that have been closed, and the number of warnings that have not been handled, as well as the processing status of each warning; Feedback collection module, used to: Collect feedback from medical staff on early warning information, including the specific content of the feedback, the time of the feedback, and the identity information of the feedback provider; Mechanism optimization module for: Based on the collected feedback information and monitoring data, the early warning mechanism is adjusted and optimized, including the adjustment of thresholds and the modification of early warning channels.

8. The liver cancer target immune drug adverse reaction active monitoring system according to claim 5, characterized in that: The report generation module has the following functions: Report conversion function: convert the generated report into specific report document formats, including PDF, Word and Excel; Add annotation function: add annotations to the generated report, including explanations and analysis of the forecast results, as well as suggestions for follow-up actions; Export and sharing function: Provides report export and sharing functions for pharmacists and doctors to export reports to local or other systems for viewing, or share reports with others for discussion and decision-making 。

Images