Detecting potential signals of adverse drug events from prescription data
作者：
Highlights：
• Detecting Adverse drug events purely from prescription data containing no additional information.
• Adapting the case-crossover study to extract samples from constructed case and control time windows.
• Using a logistic regression model to automatically evaluate multiple medicines and select the medicines with the most significant associations with the ADE via Lasso regularisation.
• Proposing improvements to take temporal effect, frequency of prescription and individual effect into account.
• As a complement to other signal detection methods for ADEs, our work can potentially strengthen global pharmacosurveillance.
摘要
•Detecting Adverse drug events purely from prescription data containing no additional information.•Adapting the case-crossover study to extract samples from constructed case and control time windows.•Using a logistic regression model to automatically evaluate multiple medicines and select the medicines with the most significant associations with the ADE via Lasso regularisation.•Proposing improvements to take temporal effect, frequency of prescription and individual effect into account.•As a complement to other signal detection methods for ADEs, our work can potentially strengthen global pharmacosurveillance.
论文关键词：Adverse drug events (ADEs),Prescription data,Logistic regression,Case-crossover
论文评审过程：Received 22 February 2019, Revised 6 February 2020, Accepted 24 February 2020, Available online 27 February 2020, Version of Record 4 March 2020.
论文官网地址：https://doi.org/10.1016/j.artmed.2020.101839
