Evaluating efficiency and statistical power of self-controlled case series and self-controlled risk interval designs in vaccine safety.

ABSTRACT

The self-controlled case series (SCCS) and self-controlled risk interval (SCRI) designs have recently become widely used in the field of post-licensure vaccine safety monitoring to detect potential elevated risks of adverse events following vaccinations. The SCRI design can be viewed as a subset of the SCCS method in that a reduced comparison time window is used for the analysis. Compared to the SCCS method, the SCRI design has less statistical power due to fewer events occurring in the shorter control interval. In this study, we derived the asymptotic relative efficiency (ARE) between these two methods to quantify this loss in power in the SCRI design. The equation is formulated as [Formula see text] (a control window-length ratio between SCRI and SCCS designs; b ratio of risk window length and control window length in the SCCS design; and [Formula see text] relative risk of exposed window to control window). According to this equation, the relative efficiency declines as the ratio of control-period length between SCRI and SCCS methods decreases, or with an increase in the relative risk [Formula see text]. We provide an example utilizing data from the Vaccine Safety Datalink (VSD) to study the potential elevated risk of febrile seizure following seasonal influenza vaccine in the 2010-2011 season.