Effect modification is defined as a differential impact of a treatment or exposure on the disease process.
- 11.2Differences in the size of associations observed in studies may or may not represent a true differential impact of a treatment or exposure.
- 11.2.1Observed differences in the size of an association are more likely to represent an actual differential effect of an exposure or treatment on the disease if:
The amount of difference in the association across subgroups is large.
The subgroups contain sufficient numbers of people and outcomes.
There is plausibility for the differential size of the association.
The differential size of the association is replicated in other studies.
Statistical measures estimate whether the differential size of an association likely exceeds chance but require careful interpretation in the context of multiple testing.
Effect modification addresses a fundamentally different concept than confounding.
The presence of effect modification on the additive scale suggests target populations that would receive the greatest benefits or harms from treatment .