Last update:
06-Jun-2005

The term "effect measure modification" was introduced by Miettinen and should be distinguished from the term "statistical interaction". The term "statistical interaction" is used to refer to a departure from the underlying form of a statistical model. However the term "statistical interaction" does not have explicit implication because there are various statistical models (additive, multiplicative and other). In epidemiology the term "effect measure modification" is used instead of the term "statistical interaction". Εffect measure modification is said to occur when the magnitude of the association between one variable and another differs according to the level of a third variable. For example, obesity increases the risk for breast cancer in postmenopausal but not premenopausal women. In this way the presence or absence of menopause modifies the effect of obesity on breast cancer. Thus menopausal status is a modifier of the effect of obesity on breast cancer. The modifier is the characteristic of individual that modifies the effect of an exposure on the incidence of disease. In this way the modifier is the characteristic that modifies the relation between exposure and disease. Εffect measure modification should be distinguished from confounding which is defined as a distortion in an effect measure introduced by an extraneous variable, occurring only in the context of a particular study. Confounding is not a characteristic of each variable. Confounding in a study is a systematic error (bias) that the investigator should locate and eliminate in order for results of the study to be valid, whereas effect measure modification is a finding to be reported rather than a bias to be avoided.

Key words: Confounder, Confounding, Effect measure modification, Interaction, Modifier.