Abstract 055: Neighborhood-Wide Association Study Of Hypertension In The Health And Retirement Study

Abstract

Background: Systematic evaluation of neighborhood factors which capture an array of characteristics -analogous to genome-wide-association studies- may identify important patterns in spatial determinants of blood pressure control. Methods: Our sample included Health and Retirement Study participants (N=13180; 58% women, 13% non-Hispanic Black, 4% Hispanic/Latino) with at least one sphygmomanometer reading taken between 2006 and 2016. Our main study outcome was at least one hypertensive blood pressure measurement over the study period. Participants were randomly assigned to either a training or test dataset. Using generalized estimating equations, we summarized multivariable associations between each of 51 standardized American Community Survey sociodemographic, housing, and income-related census tract variables (2005-09) and the period prevalence of measured hypertension. We adjusted for individual factors and accounted for multiple comparisons in the training set using the Simes significance test. Neighborhood-based factors that revealed statistically significant associations (Simes-adjusted p-value<=0.05) with hypertension in the training dataset were rerun in the test dataset to replicate findings. Lastly, in the full cohort, we evaluated main and race/ethnicity-stratified independent effects of each significant neighborhood factor on the likelihood of at least one hypertensive sphygmomanometer reading between 2006 and 2016. We hypothesized that residence in neighborhoods with worse housing or socioeconomic characteristics would be associated with a higher odds of hypertension, and that associations in the full sample would differ by racial/ethnic designation. Results: Thirty-two percent (4218 out of 13180) of participants had at least one hypertensive sphygmomanometer reading between 2006 and 2016. In the training set, two of the 51 census-tract level variables were independently associated with period prevalence of hypertension. In the full sample, we observed a lower likelihood of prevalent hypertension (OR: 0.95; 95% CI: 0.92, 0.99) among participants residing in a census tract with recent (since 2000) in-migration. A higher proportion of relatively recent (since 2000) renters in the census tract was associated with a lower hypertension prevalence (OR: 0.95; 95% CI: 0.91, 0.98). When stratified by racial/ethnic designation, these patterns were apparent among Non-Hispanic White and to some extent Hispanic/Latino but not Non-Hispanic Black participants. However, differences were not statistically significant by race or ethnicity. Conclusion: In conclusion, relatively more recent relocation to an area appears to be modestly associated with reduced prevalence of hypertension. These findings support possible differential cardiovascular health impacts of gentrification. Further investigation is needed to confirm these findings.