Weathering in Detroit: Place, Race, Ethnicity, and Poverty as Conceptually Fluctuating Social Constructs Shaping Variation in Allostatic Load.

Unlabelled: Policy Points Despite 30 years of attention to eliminating population health inequity, it remains entrenched, calling for new approaches. Targeted universalism, wellness-based local development, and Jedi Public Health approaches that are community informed, evidence based, and focused on improving everyday settings and diverse lived experiences are important policy directions. State and federal revenue transfers are necessary to mitigate the harms of austerity and assure greater equity in fiscal and population health in places like Detroit, Michigan.

Context: US population health inequity remains entrenched, despite mandates to eliminate it. To promote a public health approach of consequence in this domain, stakeholders call for moving from risk-factor epidemiology toward consideration of dynamic local variations in the physiological impacts of structured lived experience.

Methods: Using a community-based, participatory research approach, we collected and analyzed a unique data set of 239 black, white, and Mexican adults from a stratified, multistage probability sample of three Detroit, Michigan, neighborhoods. We drew venous blood, collected saliva, took anthropometric measurements, and assayed specimens to measure allostatic load (AL), an indicator of stress-mediated biological dysregulation, linking participants' AL scores and survey responses. In a series of nested Poisson models, we regressed AL on socioeconomic, psychosocial, neighborhood, and behavioral stressors to test the hypothesis that race/ethnicity and poverty-to-income ratio (PIR) are conceptually fluctuating variables whose impacts on AL are sensitive to structured lived experience.

Findings: White and Mexican Detroit participants with PIR < 1 have higher AL than counterparts nationally; black participants in Detroit and nationwide had comparable AL. Within Detroit, disparities by PIR were higher in whites than blacks, with no significant difference by PIR in Mexicans. The size of estimated effects of having PIR < 1 for whites is 58 percentage points greater than that of Mexicans and twice that of blacks.

Conclusions: Structurally rooted unobserved heterogeneity bias threatens the validity of independent main effects interpretations of associations between race/ethnicity, socioeconomic characteristics, or place and health. One-size-fits-all analytic or policy models developed from the perspective of the dominant social group insufficiently address the experiences of diverse populations in specific settings and historical moments; nor do they recognize culturally mediated protective resources residents may have developed against material and psychosocial hardship.