Collaborative management system for drug safety based on knowledge graph

Abstract

The application provides a drug product marketing license holder pharmacovigilance collaborative management system based on a knowledge graph, relates to the technical field of artificial intelligence knowledge graphs, and comprises: a pharmacovigilance capability evaluation module, which is used for evaluating the pharmacovigilance capability of a drug product marketing license holder based on a knowledge graph; a pharmacovigilance collaborative management strategy decision module, which is used for deciding the pharmacovigilance collaborative management strategy of the drug product marketing license holder according to the evaluation result; and a pharmacovigilance collaborative management strategy execution module, which is used for executing the pharmacovigilance collaborative management strategy. The application comprehensively evaluates the pharmacovigilance capability of the drug product marketing license holder through an enterprise capability knowledge graph, and formulates and executes a collaborative management strategy based on the evaluation result, so that the pharmacovigilance level is significantly improved.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence knowledge graph technology, and in particular to a drug marketing authorization holder pharmacovigilance collaborative management system based on knowledge graph. Background Technology

[0002] With the rapid development of the global healthcare industry, pharmaceuticals play a vital role in disease treatment, and their safety is directly related to public health and the sustainable development of the pharmaceutical industry. Pharmacovigilance, as a scientific system for monitoring, assessing, and preventing adverse drug reactions, plays a crucial role in ensuring medication safety.

[0003] However, the entire lifecycle regulation of pharmaceuticals encompasses multiple stages, including research and development, production, distribution, and use, involving numerous components. The lack of coordinated management across these stages presents many challenges to pharmacovigilance efforts. Traditional management models are often limited to single enterprises or specific stages, resulting in information silos and capability disparities. For example, a marketing authorization holder may possess strong safety research capabilities during the research and development phase, but may lack sufficient information feedback and risk monitoring during distribution and use; while pharmaceutical distribution units may be unable to effectively report adverse reactions due to resource and technological limitations. This fragmented management reduces efficiency and makes it difficult to address complex risks across different stages.

[0004] In recent years, technologies such as artificial intelligence and knowledge graphs have brought new opportunities to pharmacovigilance management. Knowledge graphs, through structuring complex data and relationships, have shown potential in the medical field, but their application in collaborative pharmacovigilance management for drug marketing authorization holders still needs further exploration. Existing technologies are mostly focused on the internal management of a single enterprise or the monitoring of specific processes, lacking overall integration and optimization. Furthermore, they rely on subjective judgment in enterprise capability assessments, resulting in assessments that are not comprehensive or objective enough.

[0005] To address the aforementioned issues, this invention proposes a knowledge graph-based collaborative management system for pharmacovigilance of drug marketing authorization holders. Summary of the Invention

[0006] One of the objectives of this invention is to provide a knowledge graph-based collaborative management system for pharmacovigilance of drug marketing authorization holders, in order to solve the problems pointed out in the background art.

[0007] The knowledge graph-based collaborative management system for drug marketing authorization holder pharmacovigilance provided in this invention includes:

[0008] The pharmacovigilance assessment module is used to assess the pharmacovigilance capabilities of drug marketing authorization holders based on knowledge graphs.

[0009] The pharmacovigilance collaborative management strategy decision-making module is used to determine the pharmacovigilance collaborative management strategy for drug marketing authorization holders based on the assessment results.

[0010] The pharmacovigilance collaborative management strategy execution module is used to execute the pharmacovigilance collaborative management strategy.

[0011] Optionally, the pharmacovigilance assessment module, based on a knowledge graph, assesses the pharmacovigilance capabilities of drug marketing authorization holders, including:

[0012] The pharmacovigilance capability assessment criteria for drug marketing authorization holders are determined from the knowledge graph, and the pharmacovigilance capability assessment results are determined based on the pharmacovigilance capability assessment criteria.

[0013] The pharmacovigilance capability assessment results from different companies were integrated to obtain the assessment results.

[0014] Optionally, the pharmacovigilance collaborative management strategy decision module, based on the evaluation results, determines the pharmacovigilance collaborative management strategy for the drug marketing authorization holder, including:

[0015] Identify the corresponding pharmacovigilance co-management strategy from the pharmacovigilance co-management strategy library.

[0016] Optionally, the pharmacovigilance capability assessment criteria may include at least: organizational structure and staffing information, quality management system information, document and record management information, monitoring and reporting capability information, risk control and communication information, and system and resource support information.

[0017] Optionally, the pharmacovigilance capability assessment results may include at least: the establishment and operation of the drug safety committee, the setup and performance of duties of the pharmacovigilance department, the qualifications and performance of the pharmacovigilance officer, the timeliness and completeness of information collection and reporting, the scientific nature of signal detection and risk assessment, the effectiveness and completeness of risk control measures and records, the standardization of document and record management, and the compliance of entrusted management.

[0018] Optionally, the pharmacovigilance collaborative management strategy includes at least: internal collaborative mechanisms, external collaborative mechanisms, and emergency collaborative mechanisms.

[0019] Optionally, the pharmacovigilance collaborative management strategy execution module executes the pharmacovigilance collaborative management strategy, including:

[0020] Implement pharmacovigilance collaborative management strategies through the pharmacovigilance collaborative management platform of the drug marketing authorization holder.

[0021] Optionally, the knowledge graph-based collaborative management system for drug marketing authorization holder pharmacovigilance also includes:

[0022] The knowledge graph update module is used to update the knowledge graph at preset time intervals.

[0023] Optionally, the knowledge graph-based collaborative management system for drug marketing authorization holder pharmacovigilance also includes:

[0024] The pharmacovigilance collaborative management strategy implementation effect tracking module is used to track the implementation effect of the pharmacovigilance collaborative management strategy;

[0025] The pharmacovigilance collaborative management strategy optimization module is used to optimize the pharmacovigilance collaborative management strategy based on the tracking execution effect;

[0026] The strategy relay execution module is used to relay the optimized pharmacovigilance collaborative management strategy.

[0027] Optionally, the pharmacovigilance collaborative management strategy optimization module optimizes the pharmacovigilance collaborative management strategy based on the tracked execution results, including:

[0028] Match the strategy optimization rules corresponding to the execution effect;

[0029] Based on strategy optimization rules, the collaborative management strategy for pharmacovigilance is optimized.

[0030] The present invention has achieved the following beneficial effects:

[0031] By using knowledge graphs to conduct comprehensive pharmacovigilance capability assessments of drug marketing authorization holders and developing and implementing collaborative management strategies based on the assessment results, the system significantly improves pharmacovigilance levels. This data-driven approach utilizes knowledge graph technology for accurate assessments and ensures the efficiency, synergy, and adaptability of pharmacovigilance management through the decision-making and execution of collaborative management strategies.

[0032] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0033] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0034] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0035] Figure 1This is a schematic diagram of a knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance.

[0036] Figure 2 This is another schematic diagram of the drug marketing authorization holder pharmacovigilance collaborative management system based on knowledge graph in this embodiment of the invention. Detailed Implementation

[0037] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0038] The research and development approach of this application is to use knowledge graphs to assess the pharmacovigilance capabilities of drug marketing authorization holders, and to formulate and implement collaborative management strategies based on the assessment results, thereby improving the level of pharmacovigilance.

[0039] Figure 1 A schematic diagram of a knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance is provided for embodiments of this application, such as... Figure 1 As shown, the system includes:

[0040] The pharmacovigilance assessment module 100 is used to perform step A, which assesses the pharmacovigilance capabilities of drug marketing authorization holders based on a knowledge graph. Step A specifically includes the following sub-steps:

[0041] A1. Determine the basis for assessing the pharmacovigilance capability of the drug marketing authorization holder from the knowledge graph, and determine the pharmacovigilance capability assessment result based on the basis for assessing the pharmacovigilance capability. The pharmacovigilance capability assessment criteria include at least the following: organizational structure and personnel information (whether a drug safety committee has been established, whether its responsibilities are clear and reasonable, and whether its composition meets the requirements; whether a dedicated pharmacovigilance department has been set up, and whether its responsibilities are comprehensive and clear; whether the qualifications, responsibilities, registration, and changes of the pharmacovigilance officer meet the requirements; whether the number, qualifications, and training of full-time personnel meet the requirements, etc.), quality management system information (whether pharmacovigilance quality objectives have been formulated and a quality assurance system has been established; whether quality control indicators have been formulated, and whether they are specific and measurable; whether internal audits are conducted regularly, and whether the internal audits are independent, systematic, and comprehensive, etc.), document and record management information (whether document management operating procedures have been established, and whether the system and procedures cover key activities; whether records and data are true, accurate, complete, and traceable, etc.), monitoring and reporting capability information (whether comprehensive, smooth, and effective information collection channels have been established, etc.), risk control and communication information (whether appropriate risk control measures have been taken based on the risk assessment results, etc.), and system and resource support information (whether equipment and resources sufficient for pharmacovigilance activities are provided, and whether the information system is secure and reliable, etc.). The pharmacovigilance capability assessment results shall include at least the following: the establishment and operation of the drug safety committee, the setup and performance of duties of the pharmacovigilance department, the qualifications and performance of the pharmacovigilance officer, the timeliness and completeness of information collection and reporting, the scientific nature of signal detection and risk assessment, the effectiveness and completeness of risk control measures and records, the standardization of document and record management, and the compliance of entrusted management.

[0042] The working principle of this step is to extract structured pharmacovigilance capability assessment criteria from the knowledge graph, combine them with predefined assessment standards, and systematically evaluate the pharmacovigilance capability of the drug marketing authorization holder, generating an assessment result that includes multiple capability indicators. A knowledge graph is a structured data set describing a company's pharmacovigilance capability, stored in graph form, containing nodes (such as organizational structure and quality management system) and edges (such as responsibility relationships and process dependencies). The assessment of pharmacovigilance capability is based on the following: organizational structure and personnel configuration information (obtained by reviewing the company's organizational structure documents and personnel qualification registration forms to understand the establishment of the drug safety committee, its division of responsibilities, the setup of the pharmacovigilance department, the qualifications of the person in charge, and training records of full-time personnel); quality management system information (obtained by reviewing the quality management manual and internal audit reports to understand the setting of quality objectives, the definition of control indicators, and the implementation of internal audits); document and record management information (confirmed by checking document management procedures and data records to verify the coverage of the system, the authenticity of records, and traceability); monitoring and reporting capability information (assessed by analyzing the logs and report submission records of the information collection system to evaluate the accessibility of information collection channels and the timeliness of reporting); risk control and communication information (confirmed by reviewing risk assessment reports and control measure records to verify the effectiveness of measures); and system and resource support information (assessed by checking the equipment list and information system security test reports to evaluate resource allocation and system reliability). These data are obtained through methods including: exporting relevant documents from the enterprise's internal management system (such as ERP or document management system), extracting key fields through manual review and automated tools (such as text analysis software), and generating structured data; for information collection and reporting, calling database interfaces to obtain log data, and calculating the average latency and completeness ratio of report submission; for risk control, comparing risk assessment reports with actual control measure execution records to calculate measure coverage and execution rate. Based on the above data, a weighted calculation method is used (weights are determined by regulatory requirements and industry standards, such as 30% for organizational structure and 25% for quality management system), to calculate scores for each indicator and generate pharmacovigilance capability assessment results, including the operation of the drug safety committee, the performance of pharmacovigilance department responsibilities, the performance of responsible persons, the timeliness of information collection, the scientific nature of signal detection, the effectiveness of risk control, the standardization of document management, and the compliance of entrusted management.

[0043] A2. Integrate the pharmacovigilance capability assessment results from different companies to obtain the assessment results.

[0044] The working principle of this step is to integrate the pharmacovigilance capability assessment results generated in step A1 from multiple drug marketing authorization holders. Based on a unified assessment indicator system, data standardization and cluster analysis methods are used to generate a comprehensive assessment result for the formulation of subsequent collaborative management strategies. The integration process involves extracting key indicator data (such as the drug safety committee operation score and the pharmacovigilance department's responsibility performance score) from the assessment results of each company, ensuring data consistency through data cleaning (such as removing outliers and standardizing units), and then using clustering algorithms (such as K-means) to group the companies according to their capability levels to generate a comprehensive assessment result. Key indicators are obtained through the following methods: extracting structured data (such as CSV-formatted score tables) from the assessment results database of step A1, and filtering the indicator values ​​for each enterprise using database query languages ​​(such as SQL); filling in missing data by communicating with enterprises or consulting supplementary materials (such as annual reports); standardizing data by using normalization methods (such as min-max normalization) to convert indicators of different dimensions (such as latency hours, coverage percentage) into a 0-1 range; and using cluster analysis by setting the number of clusters (such as 3 categories: high, medium, and low capabilities) and grouping them according to the Euclidean distance between indicators to generate a capability characteristic description for each group of enterprises (such as the average information collection timeliness of the high capability group being 95%). The comprehensive assessment results include each enterprise's capability score, ranking, classification label (such as high / medium / low capability), and key weaknesses (such as insufficient information collection, inadequate risk control), providing data support for collaborative management.

[0045] The pharmacovigilance collaborative management strategy decision module 200 is used to execute step B and, based on the evaluation results, decide on the pharmacovigilance collaborative management strategy for the drug marketing authorization holder. Step B specifically includes the following sub-steps:

[0046] B1. Determine the pharmacovigilance collaborative management strategy corresponding to the assessment results from the pharmacovigilance collaborative management strategy library. The pharmacovigilance collaborative management strategy shall include at least: an internal collaboration mechanism (clarifying the division of responsibilities and collaboration processes between the pharmacovigilance department and other departments, etc.), an external collaboration mechanism (clarifying the selection criteria for the entrusted party, the content of the entrustment agreement, and the requirements for periodic audits to ensure that the entrusted party has the corresponding capabilities and conditions; establishing information reporting and feedback mechanisms with pharmaceutical manufacturers, distributors, medical institutions, etc.), and an emergency collaboration mechanism (establishing an emergency response process for drug safety incidents, clarifying the responsibilities and action paths of each department in emergencies, etc.).

[0047] This step works by retrieving strategies from the pharmacovigilance collaborative management strategy library that match the comprehensive evaluation results. Based on predefined strategy mapping rules, it determines internal, external, and emergency collaboration mechanisms applicable to companies with different capability levels. The pharmacovigilance collaborative management strategy library is a database storing various collaborative management strategies, including strategy descriptions (such as internal division of responsibilities and processes, and contractor audit standards) and applicable conditions (such as capability score ranges and types of weaknesses). Internal collaboration mechanisms are obtained by querying the company's organizational structure and process documents to extract collaboration processes between the pharmacovigilance department and other departments (such as R&D and production) (such as the frequency and process of joint review of adverse reaction reports). External collaboration mechanisms are obtained by reviewing outsourcing agreements and audit reports to confirm contractor selection criteria (such as qualification requirements and resource allocation), agreement content (clearly defining responsibilities), and audit cycles (such as every 6 months). Emergency collaboration mechanisms are obtained by analyzing emergency response plan documents to obtain emergency response processes (such as event reporting deadlines and departmental responsibilities). The strategy mapping rules are based on indicators from the evaluation results (e.g., information collection timeliness <90 requires strengthening external information feedback mechanisms). Strategies are matched using a decision tree algorithm; for example, companies scoring below 80 need to strengthen internal collaboration processes. The final strategies include specific measures (e.g., establishing cross-departmental coordination teams), implementation requirements (e.g., audit frequency), and expected results (e.g., reducing reporting delays), ensuring they match the company's capabilities and weaknesses.

[0048] The pharmacovigilance collaborative management strategy execution module 300 is used to execute step C, the pharmacovigilance collaborative management strategy. Step C specifically includes the following sub-steps:

[0049] C1. Implement pharmacovigilance collaborative management strategies through the pharmacovigilance collaborative management platform of the drug marketing authorization holder.

[0050] The working principle of this step is to automatically execute the collaborative management strategy defined in step B1 through the drug marketing authorization holder's pharmacovigilance collaborative management platform, ensuring the efficient implementation of internal, external, and emergency collaboration mechanisms. The pharmacovigilance collaborative management platform is an integrated digital system that supports task allocation, process monitoring, and data feedback, including modules such as task scheduling, data interfaces, and log recording. The execution of internal collaboration mechanisms is achieved through the platform's task scheduling module, which assigns cross-departmental collaborative tasks (such as joint review meetings) to relevant departments, automatically generates task reminders and progress tracking records, and calculates task completion rates through log analysis (e.g., completion rate = number of completed tasks / total number of tasks). External collaboration mechanisms connect with the entrusted party's system through data interfaces, transmitting adverse reaction data in real time, monitoring agreement execution (e.g., checking the entrusted party's report submission delays), and generating compliance reports through the audit module. Emergency collaboration mechanisms, through the platform's event triggering module, detect adverse reaction events (e.g., monitoring event records in the database via API), automatically initiate emergency procedures (e.g., sending notifications to relevant departments), and record response times and implementation status. The platform obtains the data required for strategy execution (such as task lists and agreement terms) through database interfaces, calculates execution performance indicators (such as report submission delay and measure coverage) through log analysis, and sends the feedback data back to the strategy library to optimize subsequent strategy selection.

[0051] By using knowledge graphs to conduct comprehensive pharmacovigilance capability assessments of drug marketing authorization holders and developing and implementing collaborative management strategies based on the assessment results, the system significantly improves pharmacovigilance levels. This data-driven approach utilizes knowledge graph technology for accurate assessments and ensures the efficiency, synergy, and adaptability of pharmacovigilance management through the decision-making and execution of collaborative management strategies.

[0052] Through the implementation of the above steps, the system enables accurate assessment and collaborative management of pharmacovigilance capabilities for drug marketing authorization holders. Step A1 extracts assessment criteria from a capability knowledge graph, generates detailed assessment results, and identifies weaknesses in enterprise capabilities. Step A2 integrates assessment results from multiple enterprises to form a comprehensive capability profile, providing data support for strategy formulation. Step B1 matches targeted collaborative management strategies through a strategy library to ensure that measures match enterprise capabilities. Step C1 automates strategy execution through the collaborative management platform, improving internal collaboration efficiency, external agreement compliance, and emergency response timeliness. The overall solution significantly improves the standardization, efficiency, and safety of pharmacovigilance, reduces the risk of adverse reaction underreporting and emergency response delays, and provides technical support for collaborative management.

[0053] In some embodiments, the knowledge graph-based drug marketing authorization holder pharmacovigilance collaborative management system further includes:

[0054] The knowledge graph update module is used to execute step D, updating the knowledge graph at preset time intervals.

[0055] This module updates the knowledge graph periodically (at preset time intervals). By collecting and integrating the latest holder capability data (such as technical capabilities, resource allocation, regulatory compliance, etc.), the knowledge graph is dynamically adjusted to ensure that it reflects the real-time distribution and changes in capabilities, providing an accurate data foundation for subsequent collaborative pharmacovigilance management.

[0056] In some embodiments, the knowledge graph-based drug marketing authorization holder pharmacovigilance collaborative management system further includes:

[0057] The pharmacovigilance collaborative management strategy implementation effect tracking module is used to execute step E and track the implementation effect of the pharmacovigilance collaborative management strategy.

[0058] This module is responsible for tracking the effectiveness of pharmacovigilance collaborative management strategies. By collecting key indicator data during strategy implementation (such as adverse reaction reporting rate, processing efficiency, and collaborative response time), it evaluates the actual performance of the strategy and provides data support for optimization.

[0059] The pharmacovigilance collaborative management strategy optimization module is used to execute step F and optimize the pharmacovigilance collaborative management strategy based on the tracked execution results. Step F specifically includes the following sub-steps:

[0060] F1, Match the strategy optimization rules corresponding to the execution effect.

[0061] Based on the data provided by the pharmacovigilance collaborative management strategy execution effect tracking module, the system identifies problems or areas for improvement in the execution effect and matches them with preset optimization rules (such as adjusting collaborative processes, optimizing resource allocation, etc.).

[0062] F2. Optimize the pharmacovigilance collaborative management strategy based on strategy optimization rules.

[0063] Based on the matching rules, the system adjusts existing strategies, such as improving the coordination mechanism, optimizing task allocation, or enhancing the risk warning model, in order to improve the overall pharmacovigilance efficiency.

[0064] The strategy relay execution module is used to execute step G and relay the optimized pharmacovigilance collaborative management strategy.

[0065] This module is responsible for implementing the optimized pharmacovigilance collaborative management strategy. Successive execution ensures the optimized strategy seamlessly integrates with existing workflows, continuing to drive collaborative management of pharmacovigilance work, while also providing a foundation for the next round of tracking and optimization.

[0066] In some embodiments, the drug marketing authorization holder pharmacovigilance collaborative management system based on departmental capability knowledge graphs further includes:

[0067] The pharmacovigilance collaborative management strategy implementation effect tracking module is used to track the implementation effect of the pharmacovigilance collaborative management strategy.

[0068] The pharmacovigilance collaborative management strategy implementation effectiveness tracking module is responsible for monitoring the actual effects of the strategy implementation, collecting key indicator data (such as the extent of capability improvement and the rate of adverse event reduction), and providing a basis for the pharmacovigilance collaborative management strategy optimization module. Tracking indicators include changes in quantitative scores and improvements in weak areas.

[0069] Here is an implementation example:

[0070] Step 1: Indicator Definition

[0071] Training effectiveness: Improvement in monitoring and reporting capabilities scores for departments B and C.

[0072] Monitoring results: Percentage increase in the rate of adverse drug reaction reports for high-risk drugs.

[0073] Step 2, Data Collection:

[0074] After the training, Department B's score rose from 68 to 75, and Department C's score rose from 82 to 88.

[0075] After the monitoring strategy was implemented, the adverse reaction reporting rate increased from 50% to 85%.

[0076] A pharmacovigilance collaborative management strategy optimization module is used to optimize the pharmacovigilance collaborative management strategy based on the tracked execution results. The optimization of the pharmacovigilance collaborative management strategy based on the tracked execution results includes:

[0077] Match the strategy optimization rules corresponding to the execution effect.

[0078] Based on strategy optimization rules, the collaborative management strategy for pharmacovigilance is optimized.

[0079] The pharmacovigilance collaborative management strategy optimization module, based on the execution effect data from the pharmacovigilance collaborative management strategy execution effect tracking module, matches preset optimization rules to adjust and improve existing strategies, thereby enhancing their applicability and effectiveness. The strategy optimization rules consist of predefined conditions and adjustment schemes.

[0080] Continuing with the above implementation example:

[0081] Data shows that Department B's score improved by 7 points (less than 10 points), triggering Rule 1 (if the ability score improvement is less than 10 points, increase the training frequency). The training frequency for Department B will be adjusted from once a month to once every two weeks, and practical courses will be added.

[0082] The strategy relay execution module is used to relay the optimized pharmacovigilance collaborative management strategy.

[0083] The strategy relay execution module is responsible for implementing the optimized strategy, continuing the cycle of execution and improvement, and ensuring the continuous improvement of pharmacovigilance management.

[0084] The system described above tracks the actual effects of strategy implementation (such as improved competence scores and reduced adverse reaction reporting rates) and optimizes management strategies based on the tracking data, thereby achieving continuous improvement of the pharmacovigilance collaborative management strategy. This dynamic adjustment mechanism ensures the continued effectiveness and adaptability of pharmacovigilance management, further enhancing the overall level of pharmacovigilance.

[0085] In some embodiments, the knowledge graph-based drug marketing authorization holder pharmacovigilance collaborative management system further includes:

[0086] The module for comparing the assessment results of marketing authorization holders of similar drugs is used to perform step H, which compares the assessment results of marketing authorization holders of similar drugs in their pharmacovigilance capability assessment.

[0087] This module employs comparative analysis to cross-reference assessment results, identifying gaps, strengths, and weaknesses in the capabilities of each holder. The analysis results can provide a reference for regulatory agencies and enterprises, supporting the optimization of collaborative pharmacovigilance management and improving drug safety and regulatory efficiency.

[0088] like Figure 2 As shown, in some embodiments, the pharmacovigilance collaborative management strategy execution module 300 in the drug marketing authorization holder pharmacovigilance collaborative management system based on departmental capability knowledge graphs further includes:

[0089] The time window segmentation unit 301 is used to divide the execution cycle of the pharmacovigilance collaborative management strategy into a continuous time window sequence, and within each time window of the time window sequence, to distribute a customized management instruction set to the drug marketing authorization holder department based on the pharmacovigilance collaborative management strategy.

[0090] When dividing time windows, the entire strategy execution cycle is broken down into multiple consecutive time segments (such as weeks, ten-day periods, or months) to facilitate phased management and monitoring. The division can be based on the overall duration of the strategy, the complexity of the task, or the department's execution capabilities.

[0091] When determining the customized management instruction set, targeted task instructions are generated based on the department's role and current capability status, forming a customized management instruction set to ensure that the instructions match the department's responsibilities and resources.

[0092] Here is an implementation example:

[0093] The drug marketing authorization holders include Department A, Department B, and Department C, and the implementation cycle of the collaborative management strategy is one month (calculated as 28 days).

[0094] Step 1, Division Process:

[0095] The system divides a month into four time windows, with one window per week (7 days).

[0096] Step 2, Instruction Distribution:

[0097] Time window 1 (days 1-7):

[0098] Department A: Monitor the production data of a certain batch of drugs and submit a report 5 days in advance.

[0099] Department B: Monitor temperature control during transportation and record data daily.

[0100] Department C: Collect adverse reaction feedback and report it every 3 days.

[0101] Time window 2 (days 8-14): Adjust instructions based on feedback from time window 1, such as Department A needing to additionally test a certain indicator.

[0102] The non-standard response monitoring unit 302 is used to collect response data from various departments to the issued customized management instruction set in real time, and to determine whether there is a non-standard response to the strategy execution according to the preset judgment rules. If so, the corresponding department is designated as the target department and the current time window is designated as the target time window. The response data includes operation delay, semantic questioning frequency and execution error code.

[0103] Operational delay refers to the deviation between the time a department completes an instruction and the expected time. Semantic questioning frequency refers to the number of times a department raises questions or requests for clarification regarding the content of the instruction. Execution error codes refer to the frequency of errors or anomalies that occur during instruction execution. Non-standard reactions refer to deviations from expectations when a department executes instructions, such as excessive delays, frequent questioning, or high error rates. Judgment rules are based on preset thresholds. After identifying non-standard reactions, the problematic department (target department) and the problematic time period (target time window) are identified, providing focus for subsequent processing.

[0104] Continuing with the above implementation example:

[0105] Within time window 1, departments A, B, and C execute their respective instructions.

[0106] Step 1: Data Collection

[0107] Department A: Operational delay was 1 hour (expected 0.5 hours), semantic challenges were 0, and the error rate was 5%.

[0108] Department B: Operational delay is 3 hours (expected 2 hours), semantic challenges are 5 times (weekly), and the error rate is 15%.

[0109] Department C: Operational delay was 0.3 hours, semantic challenge occurred once, and the error rate was 2%.

[0110] Step 2, Activation Determination Rules:

[0111] Operational delay threshold: 2 hours.

[0112] Semantic questioning frequency threshold: 3 times / week.

[0113] Error code threshold: 10%.

[0114] Step 3: Determine and output the result:

[0115] Department B's delay (3 hours > 2 hours), frequency of inquiries (5 times > 3 times), and error rate (15% > 10%) all exceed the standards, and are therefore judged as non-standard responses. Department B is marked as the target department, and time window 1 is the target time window.

[0116] Comprehension ability coefficient calculation unit 303 is used to respond to non-standard reaction detection unit 302 determining the existence of non-standard reaction, and calculate the comprehension ability coefficient of the target department based on the historical execution data, personnel quality information and training data of the target department recorded in the departmental capability knowledge graph.

[0117] The comprehension ability coefficient is an indicator that quantifies a department's ability to understand and execute instructions, typically ranging from 0 to 1 or 0 to 100. Historical execution data includes past task completion rates and accuracy rates. Personnel quality information includes employee education, professional background, and work experience. Training data includes training frequency and results. A weighted calculation method is used to calculate the target department's comprehension ability coefficient based on this data.

[0118] Continuing with the above implementation example:

[0119] Department B is marked as the target department.

[0120] Step 1: Data Extraction

[0121] Historical execution data: 80% of instructions were completed in the past year.

[0122] Personnel quality information: The average educational background of employees is a bachelor's degree (score 70 / 100), and the professional matching degree is 60%.

[0123] Training data: Training is conducted once per quarter, with an average score of 75 / 100.

[0124] Step 2, Calculation:

[0125] Calculation formula: Comprehension ability coefficient = w1 × historical execution completion rate + w2 × personnel quality score + w1 × training results.

[0126] The weights are w1 = 0.4, w2 = 0.3, and w3 = 0.3.

[0127] Calculation process:

[0128] Historical execution completion rate = 0.8.

[0129]

[0130] Training score = 0.75.

[0131] Comprehension ability coefficient = 0.4 × 0.8 + 0.3 × 0.65 + 0.3 × 0.75 = 0.74.

[0132] Step 3, Output Results:

[0133] Department B's comprehension ability coefficient is 0.74.

[0134] The minimum waiting time threshold calculation unit 304 is used to calculate the minimum waiting time threshold for task execution of the target department based on the task dependency relationship and execution time data between departments in the collaborative path characterized by the distribution of customized management instruction sets to each department within the target time window.

[0135] The collaborative path refers to the dependency order of tasks among departments during the execution of various management instructions in a customized management instruction set. For example, department B can only begin transportation monitoring after department A completes monitoring. Execution time data represents the time it takes for each department to execute its corresponding management instruction. The minimum waiting time threshold is the shortest time a target department must wait for the completion of a prerequisite task before starting its own task. This threshold is calculated by subtracting a safety margin based on the average completion time and volatility of the prerequisite tasks.

[0136] Continuing with the above implementation example:

[0137] Within time window 1, the monitoring task of department A is a prerequisite for the transportation monitoring of department B.

[0138] Step 1: Data Extraction

[0139] The average completion time for monitoring tasks in Department A is 24 minutes, with a standard deviation of 2 minutes.

[0140] Task execution time for Department B: 12 ​​minutes.

[0141] Step 2, Calculation:

[0142] Calculation formula: Minimum waiting time threshold = Average completion time of preceding tasks - k × standard deviation. Where k is the safety margin coefficient, k = 2.

[0143] Calculation process: Minimum waiting time threshold = 24 - 2 × 2 = 20 minutes.

[0144] Step 3, Output Results:

[0145] Department B needs to wait 20 minutes before starting the task.

[0146] The auxiliary material generation engine unit 305 is used to generate auxiliary materials based on the customized management instruction set distributed to the target department within the target time window, the comprehension ability coefficient calculated by the comprehension ability coefficient calculation unit 303, and the minimum waiting time threshold calculated by the minimum waiting time threshold calculation unit 304, using a machine learning model; wherein, the complexity of the auxiliary materials is positively correlated with the comprehension ability coefficient and the minimum waiting time threshold, respectively; wherein, the output form of the auxiliary materials includes at least one of the following: a visual operation guide, a risk case library, or a dynamic process animation.

[0147] Supplementary materials are additional resources designed to help departments understand and execute instructions. The complexity of supplementary materials is positively correlated with the comprehension ability coefficient, meaning that a lower comprehension ability coefficient corresponds to more detailed materials. The complexity of supplementary materials is also positively correlated with the minimum waiting time threshold, meaning that a shorter minimum waiting time threshold corresponds to more concise materials.

[0148] When training the machine learning model, data related to the execution of pharmacovigilance collaborative management strategies are collected, such as departmental historical execution data, personnel quality information, training data, task dependencies, and execution time data. After cleaning, preprocessing, and labeling, deep learning frameworks such as TensorFlow or PyTorch are used to select a suitable model architecture (e.g., Transformer-based GPT or BERT for text generation, or GAN for visualization materials). Features such as comprehension ability coefficient and minimum waiting time threshold are extracted from the data. The data is divided into training, validation, and test sets for training. Hyperparameters are adjusted using the validation set, and performance is evaluated using the test set to ensure that the model can generate local auxiliary materials (such as visual operation guides) that match departmental needs. Finally, the model is optimized and deployed, and continuous monitoring and updates are performed to maintain its effectiveness.

[0149] Continuing with the above implementation example:

[0150] Department B's instruction is for transportation monitoring, with a comprehension coefficient of 0.74 and a minimum waiting time threshold of 20 minutes.

[0151] Step 1, Model Input:

[0152] Instruction set: Transport temperature monitoring.

[0153] Comprehension ability coefficient: 0.74.

[0154] Minimum waiting time threshold: 20 minutes.

[0155] Step 2, Model Output: Generate a 5-minute dynamic process animation demonstrating the temperature monitoring steps. The animation has a comprehension factor of 0.74 (relatively high), is concise yet covers key points. With a time threshold of 20 minutes, the 5-minute animation ensures departments have sufficient time to process the information.

[0156] The instruction execution correction unit 306 is used to push auxiliary materials to the target department before the target time window ends; suspend the distribution process of customized management instruction sets of related departments that have task dependencies with the target department until the instruction execution status of the target department is detected to have migrated to the completion state; monitor the instruction execution status of the target department after receiving partial auxiliary materials; and dynamically adjust the auxiliary materials according to the instruction execution status.

[0157] The purpose of pushing out materials is to help the target department overcome difficulties. The purpose of pausing instruction distribution is to ensure that dependent tasks are completed in sequence. When dynamically adjusting local auxiliary materials based on instruction execution, it is possible to further determine which type of materials the target department needs more, and adjust the auxiliary materials accordingly.

[0158] Continuing with the above implementation example:

[0159] Within time window 1, department B is the target department, and department C depends on its result.

[0160] Step 1: Material Push:

[0161] Send a 5-minute animation to Department B.

[0162] Step 2: Pause instruction distribution:

[0163] Suspend the distribution of instructions to Department C.

[0164] Step 3: Adjust as needed:

[0165] Monitoring Department B Status: If completed, restore Department C's instructions; if new issues are reported (such as temperature recording errors), adjust the supporting materials to a detailed guide.

[0166] It should be noted that pharmacovigilance involves multiple departments within drug marketing authorization holders, and collaborative management is crucial for ensuring medication safety. However, due to differences in capabilities between departments, complex task dependencies, and uncertainties in the execution process, traditional management methods struggle to monitor and respond quickly to deviations in real time, leading to poor strategy implementation, potential risks, and delays. This is particularly true in weak areas such as monitoring high-risk drugs and insufficient staffing, where precise assessment and targeted support are lacking.

[0167] To address this, the aforementioned system incorporates functional units such as time window segmentation, non-standard reaction monitoring, comprehension assessment, minimum waiting time calculation, auxiliary material generation, and instruction execution correction into the pharmacovigilance collaborative management strategy execution module. This technical solution significantly enhances the refinement, intelligence, and efficiency of pharmacovigilance collaborative management for drug marketing authorization holders. Specifically, the solution divides the strategy execution cycle into continuous time windows and distributes customized instructions to ensure the continuity and targeting of the management process. By monitoring departmental response data in real time and identifying non-standard reactions, it promptly detects and addresses execution deviations, enhancing the flexibility and adaptability of strategy execution. Based on departmental capability knowledge graphs, it calculates comprehension coefficients and generates minimum waiting time thresholds by combining task dependencies. It also utilizes machine learning models to generate auxiliary materials (such as visual operation guides or dynamic process animations) that match departmental needs, effectively improving departmental execution capabilities and collaborative efficiency. Furthermore, by pausing and resuming instruction distribution, it optimizes resource allocation and task scheduling, avoiding resource waste and execution conflicts. Ultimately, it achieves intelligent monitoring and intervention of the pharmacovigilance collaborative management process, thereby significantly improving pharmacovigilance levels and medication safety.

[0168] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A knowledge graph-based collaborative management system for pharmacovigilance of drug marketing authorization holders, characterized in that, include: The pharmacovigilance assessment module is used to assess the pharmacovigilance capabilities of drug marketing authorization holders based on knowledge graphs. The pharmacovigilance collaborative management strategy decision-making module is used to determine the pharmacovigilance collaborative management strategy for drug marketing authorization holders based on the assessment results. The pharmacovigilance collaborative management strategy execution module is used to execute the pharmacovigilance collaborative management strategy; The pharmacovigilance collaborative management strategy execution module includes: The time window segmentation unit is used to divide the execution cycle of the pharmacovigilance collaborative management strategy into a continuous sequence of time windows, and within each time window of the time window sequence, to distribute a customized set of management instructions to the drug marketing authorization holder department based on the pharmacovigilance collaborative management strategy; The non-standard response monitoring unit is used to collect response data from various departments in real time to the customized management instruction sets issued, and to determine whether there is a non-standard response to the strategy execution according to the preset judgment rules. If so, the corresponding department is designated as the target department and the current time window is designated as the target time window. The response data includes operation delay, semantic questioning frequency and execution error code. The comprehension ability coefficient calculation unit is used to respond to the non-standard reaction monitoring unit's determination of the existence of non-standard reactions. Based on the historical execution data, personnel quality information, and training data of the target department recorded in the department's capability knowledge graph, it calculates the comprehension ability coefficient of the target department. The minimum waiting time threshold calculation unit is used to calculate the minimum waiting time threshold for task execution of the target department based on the task dependency relationship and execution time data between departments in the collaborative path representing the distribution of customized management instruction sets to each department within the target time window. The auxiliary material generation engine unit is used to generate auxiliary materials based on the customized management instruction set distributed to the target department within the target time window, the comprehension coefficient calculated by the comprehension coefficient calculation unit, and the minimum waiting time threshold calculated by the minimum waiting time threshold calculation unit, using a machine learning model. The complexity of the auxiliary materials is positively correlated with the comprehension coefficient and the minimum waiting time threshold, respectively. The output form of the auxiliary materials includes at least one of the following: a visual operation guide, a risk case library, or a dynamic process animation. The instruction execution correction unit is used to push auxiliary materials to the target department before the target time window ends; suspend the distribution process of customized management instruction sets of related departments that have task dependencies on the target department until the instruction execution status of the target department is detected to have migrated to the completion state; monitor the instruction execution status of the target department after receiving partial auxiliary materials; and dynamically adjust the auxiliary materials according to the instruction execution status.

2. The knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance as described in claim 1, characterized in that, The pharmacovigilance capability assessment module, based on a knowledge graph, assesses the pharmacovigilance capability of drug marketing authorization holders, including: The pharmacovigilance capability assessment criteria for drug marketing authorization holders are determined from the knowledge graph, and the pharmacovigilance capability assessment results are determined based on the pharmacovigilance capability assessment criteria. The pharmacovigilance capability assessment results from different companies were integrated to obtain the assessment results.

3. The knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance as described in claim 1, characterized in that, The pharmacovigilance collaborative management strategy decision module, based on the evaluation results, determines the pharmacovigilance collaborative management strategy for the drug marketing authorization holder, including: Identify the corresponding pharmacovigilance co-management strategy from the pharmacovigilance co-management strategy library.

4. The knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance as described in claim 2, characterized in that, The pharmacovigilance capability assessment criteria include at least: organizational structure and personnel configuration information, quality management system information, document and record management information, monitoring and reporting capability information, risk control and communication information, and system and resource support information.

5. The knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance as described in claim 2, characterized in that, The pharmacovigilance capability assessment results shall include at least the following: the establishment and operation of the drug safety committee, the setup and performance of duties of the pharmacovigilance department, the qualifications and performance of the pharmacovigilance officer, the timeliness and completeness of information collection and reporting, the scientific nature of signal detection and risk assessment, the effectiveness and completeness of risk control measures and records, the standardization of document and record management, and the compliance of entrusted management.

6. The knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance as described in claim 2, characterized in that, The pharmacovigilance collaborative management strategy includes at least: internal collaborative mechanisms, external collaborative mechanisms, and emergency collaborative mechanisms.

7. The knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance as described in claim 1, characterized in that, The pharmacovigilance collaborative management strategy execution module executes the pharmacovigilance collaborative management strategy, including: Implement pharmacovigilance collaborative management strategies through the pharmacovigilance collaborative management platform of the drug marketing authorization holder.

8. The knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance as described in claim 1, characterized in that, Also includes: The knowledge graph update module is used to update the knowledge graph at preset time intervals.

9. The knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance as described in claim 1, characterized in that, Also includes: The pharmacovigilance collaborative management strategy implementation effect tracking module is used to track the implementation effect of the pharmacovigilance collaborative management strategy; The pharmacovigilance collaborative management strategy optimization module is used to optimize the pharmacovigilance collaborative management strategy based on the tracking execution effect; The strategy relay execution module is used to relay the optimized pharmacovigilance collaborative management strategy.

10. The knowledge graph-based collaborative management system for drug marketing authorization holders' pharmacovigilance as described in claim 9, characterized in that, The pharmacovigilance collaborative management strategy optimization module optimizes the pharmacovigilance collaborative management strategy based on the tracked execution results, including: Match the strategy optimization rules corresponding to the execution effect; Based on strategy optimization rules, the collaborative management strategy for pharmacovigilance is optimized.

Images