1023 — The role of neighborhood deprivation on all-cause mortality disparities among American Indian/Alaskan Native Veteran Health Administration Users
Lead/Presenter: Michelle Wong,
COIN - Los Angeles
All Authors: Wong MS (VA HSR&D Center for the Study of Healthcare Innovation, Implementation and Policy (CSHIIP), VA GReater Los Angeles Healthcare System), Hoggatt KJ (VA HSR&D Center for the Study of Healthcare Innovation, Implementation and Policy (CSHIIP), , VA Greater Los Angeles Healthcare System) Steers WN (VA HSR&D Center for the Study of Healthcare Innovation, Implementation and Policy (CSHIIP), VA Greater Los Angeles Healthcare System) Ziaeian B (VA HSR&D Center for the Study of Healthcare Innovation, Implementation and Policy (CSHIIP), VA Greater Los Angeles Healthcare System) Washington DL (VA HSR&D Center for the Study of Healthcare Innovation, Implementation and Policy (CSHIIP), VA Greater Los Angeles Healthcare System)
Disparities in all-cause mortality were recently identified for American Indians/Alaskan Native (AI/AN) Veteran Health Administration (VHA) users vs. non-Hispanic Whites (NHWs). Among non-Veterans, social determinants of health, particularly neighborhood deprivation, influence racial/ethnic mortality disparities. Little is known about neighborhood deprivation's role on VHA's racial/ethnic mortality disparities. We examined whether neighborhood deprivation 1) mediated all-cause mortality disparities between AI/ANs vs. NHWs, and 2) effect modified all-cause mortality disparities between AI/ANs vs NHWs.
Veteran data came from electronic health records for a national cohort of VHA users from 10/2008-9/2009, linked to the National Death Index, to assess all-cause mortality through 12/2016. We measured neighborhood deprivation from an area deprivation index (ADI), derived from 17 zip code+4-level 2000 U.S. Census socioeconomic measures. We assessed neighborhood deprivation as a mediator with Cox regression models according to Baron and Kenny procedures for mediation analysis. We assessed neighborhood deprivation as an effect modifier with a Cox regression model with an ADI-race interaction term. All models controlled for individual-level characteristics (age, sex, medical and mental health comorbidities, VA enrollment priority category, urban/rural status) and clustered standard errors at the VA facility-level.
After controlling for individual-level characteristics, AI/ANs had higher all-cause mortality (hazard ratio [HR]:1.07, 95%CI:1.03-1.11), and were more likely to live in deprived neighborhoods (odds ratio: 1.51, 95%CI:1.24-1.82). Overall, greater neighborhood deprivation was associated with increased mortality (most versus least deprived decile HR:1.26, 95%CI:1.24-1.29), however it did not mediate AI/AN all-cause mortality disparities vs. NHWs (HR % change:3.52%), nor did it modify this relationship.
Although AI/AN VHA-users lived in more deprived neighborhoods, neighborhood deprivation does not appear to mediate, nor effect modify AI/AN disparities in all-cause mortality vs. NHWs.
VHA's efforts to reduce racial/ethnic mortality disparities may require targeted efforts for AI/ANs, beyond interventions targeted for low-income groups. Other factors that better explain observed AI/AN mortality disparities should be examined, such as whether AI/ANs experience greater healthcare access barriers as they are more likely to reside in highly rural areas, have more behavioral risk factors such as hazardous alcohol use, or receive care from lower-performing VHA facilities. Cultural appropriateness of VA care should also be explored.