The Study On The Construction Of Cardio-cerebro Vascular Drug’s ADR Association Models Based On Data Mining

Adverse drug reactions (ADRs) can seriously affect people’s health, some of which areeven life-threatening, as a result, international attention to ADRs monitoring has beengrowing significantly. The total number of ADR reports received from China annually hasreached the WHO standard. Facing large amount of data collected, how to make furtheranalysis in order to evaluate the relationship between drug use, specific patients and theADRs, and provide reference and guide for clinical rational use of drug is the urgent problemto be solved for present. According to the characteristic of ADR data, the application of datamining techniques can meet the above requirements and provide evidence for rational druguse. Cardio-cerebrovascular diseases are threaten seriously people’s health and have becomethe number one killer of human death. It is important to conduct ADR monitoring ofcardio-cerebrovascular medicines, to evaluate the relationship of ADR-drug-drug users andto establish a drug safety warning mechanism in order to guarantee the safe and reasonableuse of cardio-cerebrovascular mediations, improve the treatment efficiency, and reduce thetreatment cost.In this study, based on the Guangdong and Jiangsu provinces’ ADR informationplatforms, cardio-cerebrovascular drug was selected as research subject for the high level ofcardio-cerebrovascular morbidity and mortality rates in populations. The association modelof drug and ADRs, drug users and ADRs, and drug, drug users with ADR was established,by using data mining techniques and defining the minimum support and minimumconfidence. This association model can screen the target object which has strong relationshipof drug and ADR, drug users and ADR, and ADR, drug with drug users, gain the valuablesignal products, predict and forecast the occurrence of ADR timely and effectively, detecthigh-risk groups precisely, and provide evidence for safe drug use and reduction of theadditional financial burden arising from drug-induced diseases.The innovations of this study include:(1) Apply the data mining techniques ofassociation rules and cluster analysis to the rational drug use of cardio-cerebrovasculardiseases for the first time and find the potential relationship of drug, ADR and drug usersfrom data;(2) Build ADR multidimensional data model by SQL Server2005AnalysisService (SSAS) and get seven combinations of three-dimensional association with strongrelevance and five of certain relevance;(3)Construct the ADR warning model based on12 combinations of three-dimensional association and guarantee the drug safety and reduce theadditional financial burden arising from drug-induced diseases and improve therapeuticeffect.