HW 8 MThomas

HW 8 MThomas

by Melanie Thomas -
  1. What are the different ways to account for SES in an analytic model when investigating racial/ethnic health disparities? (Hint: you should have three options). Discuss the interpretations/implications of each approach as it relates to the interest in understand health disparities by race/ethnicity. Draw a DAG for each option and reference it in your response (you do not have to post this!).

Mediation—As explored in the Lorch et al. article, SES can be a mediating factor in racial/ethnic disparities.  In this article, they use the Barron and Kenny approach to test for mediation which is a classic statistical approach.  The DAG in their paper, also illustrates how multiple mediating factors (including SES) may be “on the pathway” between race/ethnicity and the outcome of interest. 

Interaction and effect modification--- This is discussed more in the Headen et al. article.  In this example, rather than a factor on the pathway, SES is an effect modifier whereby different SES levels have different implications across the racial/ethnic groups. 

Clustering--- Although discussed as a part of MLRA in the last article, I got a little lost.  I think the implication is that both predictor variables and outcome variables are multi-level and that the most nuanced analysis will take these more sophisticated concepts into account.  

  1. Think about multilevel influences on a health outcome of interest to you. Discuss how you would study this, including measurement and analytic approaches you would use to account for exposures across multiple levels.

As discussed before, the health outcome of most interest to me is maternal depression.  The best measurement and analytic approaches would be to account for exposures across multiple levels, particularly looking at exposure of SES and stress across the lifecourse.  It would be helpful to do some detailed mediation analysis work similar to that depicted in the Lorch et al. figure to try to understand which mediating factors are contributing the most between race/ethnicity for example and disparities in maternal depression.

Re: HW 8 MThomas

by Christine Dehlendorf -

Multilevel analysis is quite complicated - and obviously the goal of this course is not to provide complete clarity but to introduce this as an approach to dealing with the complex contextual factors that we have been discussing throughout this course. Essentially, we want to make sure you all are aware of the analytic approaches you can use to move beyond the "black box" or "chronic disease" epidemiology and to do analysis, when appropriate, that takes into account the socioecological model.

In response to your discussion of SES, another option is to treat it as a "confounder". This is complex, as we discussed in class, since SES does not cause race, but there can be reasons to control for it in this way.

In terms of maternal depression, multi level effects you could consider could include neighborhood (i.e. presence of parks, maternal support), policy level factors (i.e. maternal leave), etc.