A scan statistic for identifying optimal risk windows in vaccine safety studies using self‐controlled case series design

Abstract

In examining the association between vaccines and rare adverse events after vaccination in post-licensure observational studies, it is challenging to define appropriate risk windows because pre-licensure randomized clinical trials provide little insight on the timing of specific adverse events. Past vaccine safety studies have often used pre-specified risk windows based on prior publications, biological understanding of the vaccine, and expert opinion. Recently, a data driven approach was developed to identify appropriate risk windows for vaccine safety studies that use the self-controlled case series design. This approach employs both the maximum incidence rate ratio and the linear relation between the estimated incidence rate ratio and the inverse of average person time at risk, given a specified risk window. In this paper, we present a scan statistic that can identify appropriate risk windows in vaccine safety studies using the self-controlled case series design while taking into account the dependence of time intervals within an individual and while adjusting for time-varying covariates such as age and seasonality. This approach uses the maximum likelihood ratio test based on fixed effects models, which has been used for analyzing data from self-controlled case series design in addition to conditional Poisson models.

Analyzing self‐controlled case series data when case confirmation rates are estimated from an internal validation sample

“…When planning chart reviews of relatively common adverse events, researchers often consider the number of cases and exposure status of cases. In a SCCS design, an individual’s follow-up time is partitioned into a risk interval that occurs immediately after a vaccine exposure, and control intervals occurring prior to vaccination and/or after the end of the risk interval (Farrington, 1995; Xu et al, 2011; Xu, Hambidge, McClure, Daley, & Glanz, 2013). Cases that occur during the risk interval are called exposed cases and those occurring during a control interval are called unexposed cases.…”

Bias and estimation under misspecification of the risk period in self‐controlled case series studies

“…A second method to estimate the optimal risk period and RI was subsequently proposed to circumvent the subjectivity of the /graphical approach (Xu et al ). It is based on a scan LRT statistic.…”

“…the six‐week period after MMR vaccination or 30‐day period after an infection‐related hospitalization) must be specified a priori in SCCS studies. However, the true risk period, which is referred to as the “optimal” risk period in the literature (Xu et al., ; ), is not known and must be hypothesized or explored in practice. If the goal is to unbiasedly estimate the RI of events within the optimal risk period relative to the control period, then misspecification of the risk period will lead to biased estimates.…”

“…The first is a graphical approach based on the observed linearity pattern of the estimated RIs and the maximum RI (Xu et al ). Although it can be informative, the subjectivity of the graphical approach poses challenges in practice, and a second approach using a hypothesis testing framework based on a scan likelihood ratio test (LRT) statistic was proposed (Xu et al ).…”

Integrating database knowledge and epidemiological design to improve the implementation of data mining methods that evaluate vaccine safety in large healthcare databases

“…Another consideration when defining outcomes is selecting the appropriate start and duration of the at‐risk postvaccination windows. This can be challenging even in customized protocol‐based studies where outcomes are limited to suspected, biologically plausible adverse events and data may be available to estimate timing of the reaction relative to vaccination [53–55]. As shown in Section 4, many data mining studies define a single at‐risk window (e.g., 1–42 days postvaccination) for all outcomes [41–44].…”