TY - JOUR
AU1 - VanderWeele, Tyler J.
AB - ORIGINAL ARTICLE Alternative Assumptions for the Identification of Direct and Indirect Effects a,b c Danella M. Hafeman and Tyler J. VanderWeele tween direct and indirect effects. From the above example, Abstract: The assessment of mediation is important for testing the we might wish to identify the effect of BMI on breast cancer mechanisms that explain an observed relationship between exposure risk that is mediated by increases in estrogen; this is the and disease. Several types of direct and indirect effects have been indirect effect. Alternatively, we might want to quantify the defined, broadly characterized as either controlled or natural. The identification of these effects requires a stricter set of assumptions effect of BMI on breast cancer that is not due to the pathway than those necessary for the identification of the total effect of through estrogen; this is the direct effect. exposure on disease. The particular assumptions that are required Two general categories of direct and indirect effects differ depending on the type of effect. We use an approach based on have been described in the literature: “natural” and “con- response types to derive new assumptions for the identification of 2– 8 trolled” effects. In the absence of statistical 
TI - Alternative Assumptions for the Identification of Direct and Indirect Effects
JF - Epidemiology
DO - 10.1097/EDE.0b013e3181c311b2
DA - 2011-11-01
UR - https://www.deepdyve.com/lp/wolters-kluwer-health/alternative-assumptions-for-the-identification-of-direct-and-indirect-EAVxf0ek0E
SP - 753
EP - 764
VL - 22
IS - 6
DP - DeepDyve
ER -