Reply letter: Are FDA prescription drug safety plans working? A case study of isotretinoin

摘要

We appreciate Van Bennekom and Mitchell's perspective in designing and implementing surveys to evaluate drug safety plans. The Accutane Survey is an impressive example of the potential knowledge that may be gained by these surveys. However, there are important methodological challenges in using surveys to assess FDA initiatives such as the iPledge program and more recent Risk Evaluation and Mitigation Strategies for prescription drugs.First, as the authors acknowledge in their 1995 report, surveys that measure patient behavior for a drug safety program can be subject to selection bias from differences in program adherence between those who participate and those who do not. Second, information bias due to participants failing to report non-adherence to physician instructions or patient labeling is another important validity threat.Third, surveys such as the Accutane Survey (n > 492 000) are often very costly and therefore infeasible. Finally, unless surveys are longitudinal in nature, they cannot provide serial evaluations of the impact of a policy change.