Trajectory of longitudinal lending discrimination modifies the efficacy of a skills-based intervention in stroke
Jeffrey J Wing1, Helen C S Meier2, Jenna I Rajczyk3
1College of Public Health, Division of Epidemiology, Ohio State University, Columbus, OH, USA.
Background And Purpose:
The Discharge Educational Strategies for Reduction of Vascular Events (DESERVE) study found that a skills-based intervention on systolic blood pressure (SBP) reduction in a cohort of mild/moderate strokes/transient ischemic attacks (TIA) was only effective at reducing SBP among Hispanics. We sought to better understand the differential ethnic success of DESERVE and determine if the efficacy of the intervention varied by longitudinal mortgage lending discrimination, measured by where they lived and home lending practices.
Methods:
We conducted a post hoc analysis using DESERVE, which randomized 552 stroke/TIA survivors to skills-based intervention or usual care for secondary stroke prevention. We geocoded participant addresses to census tract-level historic and present-day lending discrimination. We used a four-level neighborhood longitudinal lending discrimination variable from Home Owners' Loan Corporation redlining data and 2018 Home Mortgage Disclosure Act reports: no discrimination, growing investment, declining investment, and persistent discrimination. We modeled change in SBP by intervention status stratified by longitudinal lending discrimination category linearly with generalized estimating equations, clustering by site with inverse probability weights.
Results:
In unweighted models, the intervention was most efficacious in growing investment areas (8.65 mmHg reduction; 95 % CI: 0.48, 16.82) vs. usual care and least efficacious in declining investment areas (9.69 mmHg increase; 95 % CI:15.04, -4.34). After weighting to account for selection biases and duration of time lived in the community, conclusions were unchanged.
Conclusion:
The intervention efficacy was impacted by underlying systems of persistent disadvantage, emphasizing the need for contextual factor consideration when designing stroke recovery trials.
