Abstract — HSRD 2019
Back1172 — Mortality Correlates in the Veterans Health Administration: Interaction between Individual Socio-economic Status and Neighborhood Deprivation
Lead/Presenter: Donna Washington, COIN - Los Angeles
All Authors: Washington DL (VA HSR&D Center for the Study of Health Care Innovation, Implementation & Policy; Health Equity-QUERI Partnered Evaluation Center, Los Angeles, CA), Steers WN (VA HSR&D Center for the Study of Health Care Innovation, Implementation & Policy; Health Equity-QUERI Partnered Evaluation Center, Los Angeles, CA), , Wong MS (VA HSR&D Center for the Study of Health Care Innovation, Implementation & Policy; Health Equity-QUERI Partnered Evaluation Center, Los Angeles, CA), Ziaeian B (VA HSR&D Center for the Study of Health Care Innovation, Implementation & Policy; Health Equity-QUERI Partnered Evaluation Center, Los Angeles, CA), Hoggatt KJ (VA HSR&D Center for the Study of Health Care Innovation, Implementation & Policy; Health Equity-QUERI Partnered Evaluation Center, Los Angeles, CA)
Objectives:
Mortality disparities by individual-level socio-economic status (i-SES) and neighborhood deprivation (n-SES) have been well-established in the U.S. population. Among Veterans, low n-SES was found to be associated with increased mortality risk. However, the magnitude of i-SES effects, and whether i-SES modifies the n-SES-mortality relationship have not been examined. Our objectives were to characterize associations between i-SES and mortality, and whether i-SES modified the n-SES-mortality relationship among VA users.
Methods:
We assessed all-cause and cardiovascular mortality in Veterans using VA care 10/01/2008-9/30/2009, who had vital status and demographic data (n = 5,030,722). I-SES was categorized based on VA enrollment priority into high-income (reference group), low-income, and indeterminate-income (for Veterans with service-connected disability). N-SES was measured by the Area Deprivation Index. Deaths were ascertained through 12/31/2011. We used Cox Proportional Hazards regression to test associations between i-SES and time to mortality, accounting for demographics (age, sex, race/ethnicity, residential rurality), n-SES decile and facility-level clustering; then adding, in sequential models, i-SES by n-SES interactions, co-morbidity and primary care use. During 14,442,554 person-years at risk, 516,540 deaths were observed.
Results:
I-SES and n-SES were independently associated with all-cause mortality, with higher all-cause mortality for low-income (hazard ratio [HR] = 1.46, 95%CI 1.44-1.49) and service-connected (HR = 1.40, 95%CI = 1.37-1.42) groups, and all n-SES deciles compared with the least deprived decile (e.g., decile-10 versus decile-1 HR = 1.36, 95%CI = 1.32-1.40). I-SES modified the n-SES-mortality relationship, with increasing excess mortality for service-connected Veterans across the six most deprived n-SES deciles (interaction HRs = 1.04-1.16), but reduced excess mortality for low i-SES Veterans across 4:5 most deprived deciles (interaction HR = 0.95 for decile-6; 0.91 for decile-10). Associations remained significant although attenuated after co-morbidity and primary care use adjustment. There were similar patterns for cardiovascular mortality.
Implications:
Low-income and service-connected disability were associated with increased mortality risk. Neighborhood deprivation was an important effect modifier, with greater neighborhood deprivation attenuating elevated mortality risk for low i-SES Veterans.
Impacts:
Deprived neighborhoods appear to have a protective effect on low-income Veterans, however, we need to better understand the mechanisms (e.g., access to social resources, social support/capital). Additional alignment of VA benefits and social services with Veterans' non-medical needs may be required to address mortality risk for vulnerable Veterans.
