Olawore Oluwasolape, Stϋrmer Til, Glynn Robert J, Lund Jennifer L
Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States.
Division of Preventive Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, United States.
Am J Epidemiol. 2024 Sep 11. doi: 10.1093/aje/kwae358.
The healthy user effect is a well-recognized bias in the field of pharmacoepidemiology and can be expected to overstate the effect of a preventive intervention when comparing long term users or "adherers" to non-users. Similar to the healthy worker effect observed in occupational epidemiology, the healthy user effect can be separated into a healthy initiator effect (baseline confounding) and a healthy adherer effect (selection bias). Restriction approaches and new user designs that implicitly condition on the indication and, similar healthy behaviors or health status can often mitigate the healthy initiator effect (confounding) or healthy adherer effect (selection bias) at the start of a study. Addressing the healthy adherer effect due to continued conditioning on adherence over the duration of a study is more challenging as methods to mitigate it require the ability to predict adherence, which is often difficult using databases common in pharmacoepidemiologic research. Here, we describe the healthy user effect, with supporting examples, and describe study design approaches available to pharmacoepidemiologists to mitigate the potential for bias.
健康使用者效应是药物流行病学领域中一种广为人知的偏差，在比较长期使用者或“依从者”与非使用者时，预计会高估预防性干预措施的效果。与职业流行病学中观察到的健康工人效应类似，健康使用者效应可分为健康启动者效应（基线混杂）和健康依从者效应（选择偏倚）。限制方法和新使用者设计，即隐含地以适应症以及类似的健康行为或健康状况为条件，通常可以在研究开始时减轻健康启动者效应（混杂）或健康依从者效应（选择偏倚）。由于在研究过程中持续以依从性为条件而产生的健康依从者效应更具挑战性，因为减轻这种效应的方法需要具备预测依从性的能力，而使用药物流行病学研究中常见的数据库往往很难做到这一点。在此，我们通过支持性示例描述健康使用者效应，并描述药物流行病学家可采用的研究设计方法，以减轻偏差的可能性。
