EPI10-4: SUMMARY OF BOTH CONFOUNDING AND EFFECT MEASURE MODIFICATION |
As you can see from previous tables 1 to 4 a confounder will change the crude effect to an adjusted effect within each stratum. The stratum-based effects will however be identical or very similar and combinable in some way to provide an effect adjusted for the confounder over all the levels of this confounder. What results is the independent effect of the exposure of interest (alcohol, for example) with the effect of the confounder (in this case, tobacco) taken out of the picture.
In other cases, as shown in Tables 5 through 8, a third variable can change not only the crude effect but also the stratum specific effects that will moreover be different from each other. This type of variable is known as an effect measure modifier (EMM) as it modifies the effect from stratum to stratum. The crude effect can also be adjusted in a standard way (by direct standardisation only) to provide an adjusted effect when there is effect modification. But the usual way of showing effect modification is to show the RRs for each stratum.
Relationships between crude and stratum specific RRs | |||
---|---|---|---|
Crude RRc | Stratum 1 RR1 | Stratum 2 RR2 | All equal, hence no confounding or EMM |
Crude RRc | Stratum 1 RR1 | Stratum 2 RR2 | RR1 = RR2 ¹ RRc, hence confounding |
Crude RRc | Stratum 1 RR1 | Stratum 2 RR2 | RR1 ¹ RR2, hence EMM irrespective of how these differ from the RRc |