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.
Mediation analysis. It helps to understand if the differences among race are due to SES or due to other factors. Doing a crude analysis (without adjusting for SES) and then doing the analysis adjusting for SES, helps to understand if SES mediates the association. If in the adjusted model there is no association between race and the outcome, but in the unadjusted model, it is an association, we can conclude that the association is mediated by SES. Also, we can report the proportion of outcome mediated by SES.
Effect modification. We can also evaluate the association between another exposure (i.e. race) and the outcome are different among different levels of SES. It means that if SES modify the association between the exposure and the outcome. It is relevant because helps to understand how the SES and race interact to have an expected outcome.
Confounder. Adjusting for SES as a confounder is another way to evaluate the association between race and the outcome. However, I have doubts about this mechanism, because to be a confounder, SES should be associated with race/ethnicity and the disease. However, there is no way that SES determines race/ethnicity; it is more likely that race/ethnicity determines SES, and if this is the case, SES should be a mediator and not a confounder.
Contextual phenomenon. It is likely that people with same race/ethnicity live in the same neighborhood, and that living in the same area will account for similarities in other characteristics such as healthcare access, cultural and economic characteristics.
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.
My main interest is diabetes. In Mexico, there is evidence that as the SES increases, the proportion of a population with diabetes with an adequate glycemic control also increases, being those in the highest quintile of SES with better glycemic control (Flores-Hernandez S. et al. PLoS One. 2015; 10(7)). One of the factors that could mediate this association is the type of health insurance of the population. Those with higher socioeconomic status can have access to better health insurance. IMSS is available to those that have a formal job; people with private insurance are to those that have the opportunity to pay it (highest SES); while people who have a lower SES don't have insurance or have Seguro popular (free insurance in Mexico, subsidized by the government). My hypothesis is that private insurance and IMSS have a better healthcare quality, and therefore, better control of diabetes. Therefore, doing a mediation analysis, I would be able to understand if the differences -or what percent of the difference, among SES and diabetes control is mediated by the type of health insurance.