Abstract:
|
I will discuss the use of causal inference approaches to identify potential targets for interventions that aim to reduce or eliminate disparity, described in Jackson (2021). I will argue that the use of causal inference methods in such settings must begin with a measure of disparity that captures one's substantive beliefs about what is fair and equitable in the distribution of health and health-related outcomes, such as treatment--what I call equity value judgements. I will discuss how such judgements can be encoded into a descriptive measure of disparity and hypothetical interventions to reduce disparity via covariates that define the estimand (allowable covariates) and how the estimand can be identified with auxillary covariates that are nonetheless deemed non-allowable for defining the estimand. I will show how the related estimators reduce to existing estimators used in disparity research under certain value judgements, as well as novel ones under more meaningful judgements. Last, I will present preliminary simulation results from ongoing work (Chang et al. (2022)), based on real data, that demonstrate the importance of allowability designations under various causal structures.
|