基于BP神经网络的体外诊断试剂不良事件风险评估模型研究  

Study on Risk Assessment Model of in Vitro Diagnostic Reagent Adverse Events Based on BP Neural Network

作　　者：朱清[1] 丁静[1] 任文霞[1] 茅鸯对[1] 王雯 ZHU Qing;DING Jing;REN Wenxia;MAO Yangdui;WANG Wen(Zhejiang Pharmaceutical College, Ningbo, 315100;Center for Medical Device Evaluation in Zhejiang Province, Hangzhou, 311100)

机构地区：[1]浙江医药高等专科学校,宁波市315100 [2]浙江省医疗器械审评中心,杭州市311100

出　　处：《中国医疗器械杂志》2019年第2期136-139,共4页Chinese Journal of Medical Instrumentation

基　　金：国家药品监督管理局药品评价中心项目(诊断类医疗器械不良事件监测模式研究)

摘　　要：目的改进体外诊断试剂不良事件监测流程和手段,提高不良事件上报数量和质量,减少监管部门工作量,保障体外诊断试剂的安全有效。方法基于BP神经网络的前置过滤风险评估系统对体外诊断试剂不良事件进行初步评价,监管部门根据评价结果采取相应的应对措施。结果 BP神经网络通过历史数据学习对不良事件所作的风险评价结果与专家组的基本一致。结论利用BP神经网络可对不良事件进行风险评估,实现不良事件的风险信号聚集。Objective To modify the monitoring process and means of adverse events in vitro diagnostic reagents, improve the quantity and quality of adverse events reported in vitro, and reduce the workload of regulatory authorities, eventually ensure the safety and effectiveness of in vitro diagnostic reagents. Methods The pre-filtering risk assessment system based on BP neural network was used to evaluate the adverse events of in vitro diagnostic reagents. According to the evaluation results, the administrative supervision departments took corresponding countermeasures. Results The BP neural network learned the historical data, and the risk evaluation results of the adverse events were basically consistent with the expert group. Conclusion BP neural network can be used to evaluate the risk of adverse events and achieve risk signal aggregation of adverse events.

关 键 词：体外诊断试剂 不良事件 前置过滤 BP神经网络 风险评价 

分 类 号：TP301[自动化与计算机技术—计算机系统结构] R95[自动化与计算机技术—计算机科学与技术]