机器学习方法在预测麻精药品不合理使用风险中的应用现状和思考

Application of machine learning methods in predicting the risk of irrational use of narcotic and psychotropic drugs:current status and considerations

麻醉药品、精神药品不合理用药在欧美国家已造成了严重的公共卫生问题,评估麻精药品的滥用及其他不合理用药模式风险、监督麻精药品全流程合理合规使用是监管工作的重点难点。近年来,国外借助真实世界数据,采用机器学习方法构建预测模型以快速识别药物滥用和药物使用障碍,以及预测药物依赖、持续使用等不合理使用模式和不良反应等研究日益增多,然而我国学者对类似研究范式关注仍较少。通过梳理麻精药品预测模型研究现状,集中关注阿片类药物用药风险预测相关研究,概括研究场景和研究设计要点,提出对模型转化和我国监管重点的思考,以期为将机器学习用于中国麻精药品监管领域提供思路。

The irrational use of narcotic drugs and psychotropic drugs has let to significant public health issues in Europe and the United States.It is a key challenge in the regulatory work to assess the risk of drug abuse and other irrational use pattern,and to supervise the entire process of the use of narcotic and psychotropic drugs.Over recent years,an increasing number of studies oversea have used machine learning methods to build predictive models to rapidly identify drug abuse and drug use disorders,predict drug dependency,persistent use and other irrational use patterns and adverse effects using real-world data,while Chinese scholars still pay less attention to similar research paradigms.This paper compares the status of research on narcotic and psychotropic drug prediction models,mainly focuses on the related research of opioid drug risk prediction,summarizes the research scenarios and key points of research design,as well as presents considerations on model transformation and regulatory priorities for China,aiming to provide suggestions for the use of machine learning in the field of narcotic and psychotropic drug regulation in China.