摘要
目的建立药物不良反应监测模拟数据,为研究药物不良反应信号检测方法提供模拟性实验的平台。方法参照蒙特卡罗方法基本思想,基于不相称性测定理论,采用随机模拟的方法对药物不良反应监测数据进行模拟。结果本文随机模拟方法可以灵活设置记录总数、药物种数、不良反应种数、数据构成、目标药物-事件组合方式及出现频数、时间序列等。该方法产生的海量模拟数据含有联合用药的信息,数据中的目标药物-事件组合在不同时段出现的频率可以呈一定趋势波动。结论基于不相称性测定理论的数据模拟方法仿真性、扩展性和可控制性强,能够广泛适用于各种信号检测方法,对研究各种方法的灵敏度和特异度、优点和缺点以及应用的可行性和实用性等具有十分重要的意义。
Objective To generate high quality simulation data of adverse drug reaction surveillance, which could be used to explore on the method applied to detect adverse drug reaction signal. Methods In accordance with Monte Carlo method and measures of disproportionality theory, simulate the adverse drug reaction surveillance data by the random mode. Results The method discussed in this article can flexibly set the total recorders of the simulation data, the number of drugs and adverse drug reactions, the constitution of the simulation data, the combination mode and frequencies of the objective events and the serial of time. The simulation data generated contains the information of drug - drug interaction. In addition, the frequencies of the objective events in the simulation data can undulate by the serial of time. Conclusion The simulation method based on the measures of disproportionality theory can be extended, and the simulation data generated can be applied very abroad. It is an important and practicable method to provide simulation data to evaluate the sensitivity and specificity, merit and shortcoming, feasibility and practicability of various methods of detecting adverse drug reaction signal.
出处
《中国卫生统计》
CSCD
北大核心
2008年第4期343-346,共4页
Chinese Journal of Health Statistics
关键词
药物不良反应
数据模拟
数据挖掘
Adverse Drug Reaction
Data Simulation
Data Mining
