Effect Modification and Interaction Terms: It Takes Two to Tango.

Abstract

In this Investigators' Corner I look more deeply into the previously discussed phenomenon of effect modification. I revisit an explanation and examples of the phenomenon and then examine how to account for it statistically. Specifically, I show, in detail, how to write a regression equation that includes interaction terms that account for the effect modification. Finally, I look at interpretation of regression coefficients both with and without the presence of effect modification, and the associated interaction terms.