| Adverse Drug Reaction(ADR) monitoring is a cornerstone of drug safety post-marketing surveillance.Presently,Spontaneous Reporting System(SRS),Intensive Hospital Monitoring(IHM),Prescription-Event Monitoring(PEM) are the common monitoring methods.While SRS is the most popular method used to collect adverse drug reaction information around the world including China.Whether a signal could be generated accurately depends on the quality of ADR reports.In this study,we evaluate the quality of ADR reports in Shanghai according to The Specification Ranking Standard of Adverse Drug Reactions/Events,which was issued by the State Food and Drug Administration(SFDA),to get a general idea about it.A total number of 32752 ADR reports were collected by ADR Monitoring Center of Shanghai from 2003 to 2007.There was already an average of 650 ADR reports per million inhabitants in 2006,61.94%of which were collected by practitioners,and 3.1% of the reports were serious adverse drug reactions.We apply logistic regression with stepwise method to analyze the risk factors,odds ratio and 95%CI of the quality of ADR reports,and get the result as follows:unexpected serious ADR(OR=2.187,95%CI: 1.378～3.472),unexpected common ADR(OR=2.130,95%CI:1.514～2.997),known serous ADR(OR=5.316,95%CI:4.152～6.805),pharmaceutical company(OR=35.743, 95%CI:28.270～45.191),military hospital(OR=21.051,95%CI:28.270～45.191), monitoring agency(OR=21.051,95%CI:28.270～45.191),individual(OR=11.165, 95%CI:2.333～53.434),others(OR=2.219,95%CI:1.035～4.755).The ADR reporting rate per million inhabitants annually from practitioners,in Shanghai satisfies the requirement of WHO standard,but serious ADR reporting rate is under the minimun requirement(30%) of WHO standard.Causal inference and its reliability to signal assesment is a key and difficult step of ADR monitoring.At present the main methods for evaluating are as follows:①.global inspection②.reasoning method③.Score calculation method④probability method. Comphrensive analysis reasoning method is adopted by SFDA.In this study,we choose the result of ADR report evaluated by SFDA ADR monitoring center as authority standard and reports from Shanghai ADR monitoring center as resource on the basis of this method.Particle swarm optimization algorithm was used to find weights for 16 answer options of 5 questions,which were a part of the ADR report.Finally a scoring system was established according to the weights.Particle swarm optimization algorithm is a new Evolutionary Algorithm developed in recent years.It starts from random solution and gets optimization by iteration,which has the advantages of easy implement,high precision and fast convergence,et al.Test of scoring system based on the particle swarm optimization algorithm performed Shanghai ADR data shows that:①228 ADR cases were are classified as 'definite' by our algorithm from 285 ADR cases which had 'definite' causality identified by experts from SFDA,74 ADR cases were classified as 'indefinite' by our algorithm from 105 ADR cases which had 'indefinite' causality identified by experts from SFDA.The sensitivity and specificity were 80.0%and 70.5%,respectively.②Comparison of the scoring system with comprehensive analysis reasoning method showed that the curve area under scoring system was 0.966,standard error was 0.008;the curve area under the reasoning method was 0.876,standard error was 0.017;the asymptotic significance were all less than 0.05,their 95%confidence interval were (0.951,0.981) and(0.842,0.909).Scoring system appears to be slightly better than the comprehensive analysis reasoning method.The ADR cases which were not classified by SFDA were calculated causality probability with scoring system,the results were 170 cases were "definite" and 1275 cases were "indefinite".The scoring system is proved to be scientific and simple in this study,and could be a useful ADR screening tool in the future.It can not only help authorities and pharmaceutical companies to generate alerting ADR signal,but also help to assist clinicians to identify rare ADR.
|
PDF Full Text Request