Associations Between Social Drivers of Health and Intervention Reach During Implementation of a Supervised Walking Program.
To examine associations between social drivers of health and veteran-level reach of STRIDE, a supervised walking program implemented in the Veterans Health Administration (VA).
We included 2527 patients across 5 facilities. We tested whether housing insecurity, rurality, race, and neighborhood deprivation were associated with reach, defined as any STRIDE walk during hospitalization.
We used the first 6 months of postimplementation data from a stepped-wedge implementation trial.
During the implementation period, 197 veterans (7.8%) received at least 1 STRIDE walk. In unadjusted models, patients residing rurally (vs. urban) and those who were White, non-Hispanic/Latino (NH) (vs. Black NH) were significantly more likely to receive ≥1 STRIDE walk. Unadjusted models showed large site differences in STRIDE reach, ranging from 3% to 26% probability by site. Reach differences coincided somewhat with facility-level racial composition: the lowest reach was observed at a site with a majority of Black NH admitted patients, whereas the highest reach was found at a site with a majority of White NH patients. In adjusted models, patient-level rurality remained associated with greater reach (probability 9% for rural veterans vs. 6% for urban veterans), but race was no longer associated (probability 6% among White NH veterans vs. 7% among Black NH veterans).
Patient-level rurality was consistently positively associated with implementation outcomes. Patient-level race was associated with implementation outcomes in unadjusted but not adjusted models. Secondary analyses in implementation trials may help assess how social drivers of health are associated with implementation outcomes.
We included 2527 patients across 5 facilities. We tested whether housing insecurity, rurality, race, and neighborhood deprivation were associated with reach, defined as any STRIDE walk during hospitalization.
We used the first 6 months of postimplementation data from a stepped-wedge implementation trial.
During the implementation period, 197 veterans (7.8%) received at least 1 STRIDE walk. In unadjusted models, patients residing rurally (vs. urban) and those who were White, non-Hispanic/Latino (NH) (vs. Black NH) were significantly more likely to receive ≥1 STRIDE walk. Unadjusted models showed large site differences in STRIDE reach, ranging from 3% to 26% probability by site. Reach differences coincided somewhat with facility-level racial composition: the lowest reach was observed at a site with a majority of Black NH admitted patients, whereas the highest reach was found at a site with a majority of White NH patients. In adjusted models, patient-level rurality remained associated with greater reach (probability 9% for rural veterans vs. 6% for urban veterans), but race was no longer associated (probability 6% among White NH veterans vs. 7% among Black NH veterans).
Patient-level rurality was consistently positively associated with implementation outcomes. Patient-level race was associated with implementation outcomes in unadjusted but not adjusted models. Secondary analyses in implementation trials may help assess how social drivers of health are associated with implementation outcomes.
Authors
Wilson Wilson, Das Das, Byrd Byrd, Kappler Kappler, Stechuchak Stechuchak, Thomas Thomas, Beaver Beaver, Olsen Olsen, Johnson Johnson, Van Houtven Van Houtven, Hastings Hastings
View on Pubmed