1. What are 3 different ways to account for SES in an analytic model when investigating racial/ethnic health disparities? (describe a fourth for extra bonus points!). Briefly discuss the interpretations/implications of each approach as it relates to the interest in understand health disparities by race/ethnicity.
a) Mediation testing: Using sequential logistic regression to add mediating variables to model and determining % that the variable contributes to the point estimate. In research that shows a difference in health outcome by race/ethnicity, this is a way to study variables that may explain the health disparity.
b) Clustering: Partitioning between contextual “neighborhood” level component and individual level component. This is a method to quantify “contextual phenomena,” wherein similarities are greater in individuals within certain groups (i.e., SBP by neighborhood, as in Merlo) or clusters based on “composition” (i.e., fetal death in pregnancies nested within one woman, as in Headen).
c) Effect measure modification: Stratifying outcomes by subgroups of the study population to determine if the predictor has variable outcomes by subgroup. For example, the Headen article gives an example of the association between race/ethnicity and gestational weight gain being modified by classification of pre-pregnancy BMI.
2. Describe a potential effect modifier, mediator, or contextual variable (for definition of contextual variable, see first page of option Merlo reading) for an association of interest to you and relevant to health disparities. For example, for investigating the association between SES and maternal mortality, I might be interested in the contextual variable of exposure to violence in the neighborhood. Describe how you would study whether this relationship exists.
In studying the effect of SES on unintended teen pregnancy, one could stratify by whether the teen was born to teen mother herself (i.e., study if the effect of SES on unintended teen pregnancy is modified by the teen pregnancy being a second-generation teen pregnancy). The child of a teenage mother may be better equipped to protect herself against early unintended pregnancy because of more anticipatory guidance, or she may be more likely to have early unintended pregnancy because of fewer resources or different social norms. One could also assess the mean age of at first pregnancy among neighborhoods or high school populations to assess how the “contextual variable” affects the outcome.