Incident User Design in Comparative Effectiveness Research.

摘要

When conducting comparative effectiveness research with cohort studies or registries, some investigators restrict enrollment to patients who were incident users of an intervention; other investigators enroll all patients who used an intervention. An incident user design follows patients from the day that they started an intervention, which may reduce biases, such as confounding. But an incident user design also reduces the precision of comparative effectiveness estimates. The investigators need to weigh the tradeoffs between bias and precision when they are considering a new-user design. Our commentary considers the following question: How important is it for investigators to follow patients from the day they started treatment with the study interventions. Our objective is to start a dialog on the value of the incident user design in comparative effectiveness research. We reviewed published case studies in which investigators had used the incident user design or alternative design. We also reviewed methods papers on the incident user design. Our commentary was informed by expert opinion, not systematic evidence.