InteractionR: An R package for full reporting of effect modification and interaction

Abstract

Effect modification and/or Interaction are frequently assessed in epidemiological research. However, in most cases, authors do not present sufficient information for the readers to fully assess the extent and significance of interaction on both additive and multiplicative scale. Also, due to being readily available in most software, the delta method has proliferated in the literature for the estimation of confidence intervals (CIs) for measures of additive interactions; despite its well documented poor performance compared to alternative methods. We introduce interactionR, an open-source R package with user-friendly functions that ensures full reporting of effect modification or interaction based on recommended guidelines. In addition to the simple asymptotic delta method, the package also allows for estimation of CIs for additive interaction measures using the variance recovery and percentile bootstrapping methods.