Research On The Visualization System Construction Of Drug Safety Knowledge Towards Decision Support

Objective:Drug use is directly related to the health protection and support of the people,and ensuring pharmacovigilance is a major public safety issue.Therefore,Therefore,how to establish an effective drug safety evaluation framework to reduce the occurrence of adverse drug reactions has become the most concerned issue in the world.Currently,collecting and analyzing adverse drug reaction reports for safety evaluation is an important means of drug safety assessment after the drug hit the market.By applying data visualization techniques,reseachers could discover the hidden connections in the raw data.It could benefit people's understanding of high-dimensional and large-scale drug data.During the 12 th Five-Year Plan,China has planed and implemented the drug safety detection,assessment and monitoring system.As for the next five-year plan,China proposes that it is necessary to further improve the research and evaluation on the drug safety,constructthe safety detection platform and database and develop medical databases,etc..Therefore,it is of vital importance to explore and exploit the multi-perspective and data mining techniques and related surveillance systems,etc..As a result,this thesis will focus on the knowledge visualization on the pharmacovigilance under the prospect of medical informatic.By appling corresponding techniques,we conduct in-depth research on data extraction,information construction and knowledge discovery on pharmacovigilance.Our research could provide efficient support for the medical regulatory bureau for policy decision.Moreover,it will improve the efficiency during the communication among researchers,doctors and the general public.Methods:In this thesis,we follow the knowledge conversion process,specifically DIKW structure,to construct the paradigm of pharmacovigilance.Based on the data from Vaccine Adverse Event Reporting System(VAERS),we group the medical adverse effect reactions into different categories.In this process,we adopted to rules from both Medical Dictionary for Regulatory Activities(MedDRA)and the National Center for Biomedical Ontology(NCBO).Then we apply the XAMPP framework,consisting Apache,MySQL,PHP etc.,to construct a drug safety knowledge visualization system emphasizing on the user experience.Finally,we verify our research on the HPV vaccine database.By combing various statistical techniques,such as the Poisson regression,Pearson correlation and clustering analysis,we analyze the adverse drug reactions in HPV vaccines given different ages and genders and visualized the relevant results correspondingly.Results:Following the development of medicine information,this study summarizes the flow of the information in the study of adverse drug reaction data through the DIKW model,construct the visualization model of drug safety by using the data from VARES,Preferred Term(PT)and System Organ Class(SOC).Considering the 5 elements of user experience,we attempt to predict the vaccine safety from multiple prospects,establish the visualization system on the pharmacovigilance and complete data visualization,hot spot detection and geographic regional analysis on adverse drug reaction.Finally,we extracted 26023 adverse reactions from the HPV vaccine datasets.Based on this subset,we conducted a safety retrospective study of the symptoms of HPV vaccine adverse reactions and revealed the distribution of adverse reaction among age group and gender difference.The age difference was statistically significant among all SOCs except for 5 SOCs.Among these,we clustered the adverse reaction into nervous system diseases,mental illnesses,musculoskeletal,connective tissue diseases,and gastrointestinal diseases.This result demonstrates the connection between various adverse reactions and it helps to recognize that the benefits of HPV vaccination outweigh the potential risks.Conclusion:Through the analysis of the vaccine safety issues from multi-prospects,this study completes the data-information-knowledge transformation process,which laid the foundation for wisdom.(1)Through the research on the visualization model of pharmacovigilance,the data-information-knowledge-intelligence transformation process not only conforms to the drug safety visualization management,but also is the evolution process of the mainstream theoretical basis.Data visualization,information visualization,and knowledge visualization accelerate the process of information flow among them.The mutual transformation is aimed at the knowledge generation and sharing of vaccinated population,clinically relevant researcher,vaccine research researcher and drug surveillance managers.(2)The research achieves the transformation,cleaning,hierarchical classification,mapping and construction of drug safety level database with respect of VAERS data.This significantly improved the quality of drug safety data,and thus,provided the guarantee of data mining on pharmacovigilance.Meanwhile,this work constructed both the PT level application database and the SOC level application database respectively,providing a well-structured data source for the subsequent construction of visualization system.(3)Through different information technologies,such as informatics,medicine,statistics,human-computer interaction and other interdisciplinary knowledge,the drug safety knowledge is extracted and shared using the pharmacovigilance visualization system.It satisfies the knowledge and utilization of drug safety knowledge among people in different fields.This visualization system makes information easy to summarize and digest,accelerates the transformation between information and knowledge,and grasps the overall trend of various vaccine products.It could provide direction for the following research on the drug safety.(4)Last but not the least,the statistical relationship between the inoculation of HPV vaccine and some diseases was found after studying the relevant statistical methods for the HPV vaccine,of which provided a future direction for subsequent clinical research,drug research and public health research correspondingly.The knowledge discovery of drug safety at different levels would contribute to the decision making in drug safety management.Through the overall process,it has created a paradigm to transform the knowledge of drug safety into wisdom and provides an effective framework for the future research on the safety of other types of drugs.