Problem 5 - interaction
Learning goals:
1) What is interaction (subtypes / how to determine / causes of interaction)
The term interaction describes a situation in which two or more risk factors modify the effect of each
other with regard to the occurrence or a level of a given outcome. It is also known as effect
modification.
For dichotomous variables, interaction means that the effect of the exposure on the outcome differs
depending on whether another variable is present.
Positive interaction / synergistic = if interaction exist and the presence of the effect modifier
strengthens (accentuates) the effect of the exposure of interest.
Negative interaction / antagonistic = if the presence of the effect modifier diminishes or eliminates
the effect of the exposure of interest.
In the case of continuous variables, interaction means that the effect of exposure on outcome
depends on the level of another variable.
A minimum of three factors is needed for interaction to occur:
- Risk factor / exposure = A
- Outcome = Y
- Potential effect modifier = Z
The term homogeneity indicates that the effects of risk factor A are homogenous are similar in strata
formed by factor Z. Heterogeneity of effects implies that these effects are dissimilar.
Interaction can be defined in two different ways:
- Definition based on homogeneity or heterogeneity of effects: interaction occurs when the
effect of risk factor A on the risk of an outcome Y is not homogeneous in strata formed by
third variable Z.
- Definition based on the comparison between observed and expected joint effects of risk
factor A and third variable Z: interaction occurs when the observed joint effect of A and Z
differs from the expected joint effect based on their independent effects.
Ways to measure the interaction:
- Additive model = attributable risk
- Multiplicative model = relative risk (RR)
Causes of interaction:
- Nature
- Chance; due to sampling
- Confounding
- Other bias sources
, Problem 5 - interaction
ASSESSMENT OF HOMOGENEITY EFFECTS
Additive interaction is considered to be present when the attributable risk in those exposed to factor
A varies as a function of a third variable Z. The easiest way to evaluate interaction in this instance is
to calculate the attributable risk for those exposed to risk factor A for each stratum defined by the
level of the potential effect modifier Z.
Multiplicative interaction is considered to be present when the relative difference (ratio) in the risk of
outcome Y between subjects exposed and non-exposed to a risk factor A differs as a function of a
third variable Z.
COMPARING OBSERVED AND EXPECTED JOINT EFFECTS
An alternative definition of interaction is when the observed joint effect of A and Z differs from the
expected joint effect. The expected joint effect can be estimated by assuming that the effects of A
and Z are independent.
If there is no interaction when exposure occurs in both the factors, the observed joint effect
is expected to be the sum of the independent effects.
o Sum = combined effects of A and Z in either absolute or relative terms.
Learning goals:
1) What is interaction (subtypes / how to determine / causes of interaction)
The term interaction describes a situation in which two or more risk factors modify the effect of each
other with regard to the occurrence or a level of a given outcome. It is also known as effect
modification.
For dichotomous variables, interaction means that the effect of the exposure on the outcome differs
depending on whether another variable is present.
Positive interaction / synergistic = if interaction exist and the presence of the effect modifier
strengthens (accentuates) the effect of the exposure of interest.
Negative interaction / antagonistic = if the presence of the effect modifier diminishes or eliminates
the effect of the exposure of interest.
In the case of continuous variables, interaction means that the effect of exposure on outcome
depends on the level of another variable.
A minimum of three factors is needed for interaction to occur:
- Risk factor / exposure = A
- Outcome = Y
- Potential effect modifier = Z
The term homogeneity indicates that the effects of risk factor A are homogenous are similar in strata
formed by factor Z. Heterogeneity of effects implies that these effects are dissimilar.
Interaction can be defined in two different ways:
- Definition based on homogeneity or heterogeneity of effects: interaction occurs when the
effect of risk factor A on the risk of an outcome Y is not homogeneous in strata formed by
third variable Z.
- Definition based on the comparison between observed and expected joint effects of risk
factor A and third variable Z: interaction occurs when the observed joint effect of A and Z
differs from the expected joint effect based on their independent effects.
Ways to measure the interaction:
- Additive model = attributable risk
- Multiplicative model = relative risk (RR)
Causes of interaction:
- Nature
- Chance; due to sampling
- Confounding
- Other bias sources
, Problem 5 - interaction
ASSESSMENT OF HOMOGENEITY EFFECTS
Additive interaction is considered to be present when the attributable risk in those exposed to factor
A varies as a function of a third variable Z. The easiest way to evaluate interaction in this instance is
to calculate the attributable risk for those exposed to risk factor A for each stratum defined by the
level of the potential effect modifier Z.
Multiplicative interaction is considered to be present when the relative difference (ratio) in the risk of
outcome Y between subjects exposed and non-exposed to a risk factor A differs as a function of a
third variable Z.
COMPARING OBSERVED AND EXPECTED JOINT EFFECTS
An alternative definition of interaction is when the observed joint effect of A and Z differs from the
expected joint effect. The expected joint effect can be estimated by assuming that the effects of A
and Z are independent.
If there is no interaction when exposure occurs in both the factors, the observed joint effect
is expected to be the sum of the independent effects.
o Sum = combined effects of A and Z in either absolute or relative terms.
