Recent Changes in Health Insurance Coverage for Urban and Rural Veterans: Evidence from the First Year of the Affordable Care Act

Recent Changes in Health Insurance Coverage for Urban and Rural Veterans: Evidence from the First... Abstract Introduction Prior to the Affordable Care Act, as many as 1.3 million veterans lacked health insurance. With the passage of the Affordable Care Act, veterans now have new pathways to coverage through Medicaid expansion in those states that chose to expand Medicaid and through private coverage options offered through the Health Insurance Marketplace. We examined the impact of the ACA on health insurance coverage for veterans in expansion and non-expansion states and for urban and rural veterans. Methods We examined changes in veterans’ health insurance coverage following the first year of the ACA, focusing on whether they lived in an urban or rural area and whether they live in a Medicaid expansion state. We used data on approximately 200,000 non-elderly community-dwelling veterans, obtained from the 2013–2014 American Community Survey and estimated differences in the adjusted probability of being uninsured between 2013 and 2014 for both urban and rural areas. Adjusted probabilities were computed by fitting logistic regressions controlling for age, gender, race, marital status, poverty status, education, and employment. Results There were an estimated 10.1 million U.S. non-elderly veterans in 2013; 82% lived in predominantly urban areas (8.3 million), and the remaining 18% (1.8 million) lived in predominately rural areas. Most veterans lived in the South (43.6%), and rural veterans were more likely to be Southerners than their urban counterparts. On every marker of economic well-being, rural veterans fared worse than urban veterans. They had a statistically significant higher chance of having incomes below 138% of FPG (20.0% versus 17.0%), of being out of the labor force (29.1% versus 23.0%), and of having no more than a high school education (39.6% versus 28.8%). Rural veterans were also more likely to experience at least one functional limitation. Overall, veterans in Medicaid expansion states experienced a significantly larger increase in insurance compared to veterans living in non-expansion states. For rural veterans in Medicaid expansion states, the increase in insurance was 3.5 percentage points, compared with 1.2 percentage points in non-expansion states. Conclusion Our analysis found a substantial 24% relative decline in the rate of uninsurance for U.S. Veterans, from 9.3 to 7.1% between 2013 and 2014. We found that coverage gains in rural areas were due to gains in Medicaid and individual market coverage. Residence in a Medicaid expansion state was particularly influential for rural veterans – the increase in the insured rate was three times larger in Medicaid expansion states versus non-expansion states. The ACA has had a positive and significant impact on the ability of U.S. Veterans to obtain health insurance coverage specifically for low-income veterans living in rural areas. The poverty rate among Veterans is rising and is particularly an issue for the more recent Gulf War veterans. Providing affordable and accessible health insurance options is part of our commitment to those who have served our country. Our analysis also presents yet another reason for the 17 non-expansion states to consider a Medicaid expansion. INTRODUCTION There are more veterans living in the USA today than at any other time since WWII.1 While the Department of Veterans’ Affairs (VA) plays an important role in providing health care services to veterans through the Veterans Health Administration (VHA), 55% of veterans rely on care obtained through employer-sponsored health insurance or public programs like Medicare and Medicaid.2 In addition, approximately 1.3 million non-elderly veterans lacked comprehensive health insurance and did not access VHA health services in 2010.3 Given that many veterans are covered outside of the VHA while others lack any coverage at all, recent health reform initiatives have the potential to substantially affect veterans’ access to health care. The passage of the Affordable Care Act (ACA) in 2010 created new coverage options for veterans. Eligible veterans may now obtain health insurance coverage through newly formed Health Insurance Marketplaces with tax credits for those that qualify. The ACA also allowed states to expand their Medicaid program to non-elderly adults with incomes below 138% of the poverty level (or $16,100 of annual income for a single adult in 2014).4 As of early 2016, 31 states and the District of Columbia expanded Medicaid under the Affordable Care Act, 2 states are considering expansion, and 17 states have no plans to expand.5 Although the full implementation of the ACA is only in its third year (as of 2016), recent research suggests that, in the general population, it is having substantial impacts on health insurance coverage, access to care, and self-reported health.6–9 Recent estimates from the National Health Interview Survey (NHIS) suggest that between 2013 and the first three quarters of 2015, the number of uninsured persons declined by 16 million.9 Reductions in the uninsured have been the largest for people with lower incomes, Hispanics, and residents of states that opted to expand Medicaid. There is little evidence to date regarding changes in coverage for rural versus urban populations, but one recent study suggests that individuals in rural areas were less likely to sign up for coverage in the Health Insurance Marketplace compared to people in urban areas.10 Reductions in the uninsured have been accompanied by increases in the number of non- elderly adults that report having a regular physician, a decrease in the fraction of non-elderly adults that needed but could not afford care, and a decrease in the number that report being in fair or poor health.6 To date, there is limited evidence of substantial declines in labor force participation,11 and the rate of employer-sponsored insurance has also remained steady, suggesting that consumers have not substituted private coverage for new public coverage options (i.e., “crowd-out”).9 Given their distinct demographic profile, health needs, and the dedicated health systems serving them, veterans are likely to be affected by the ACA in unique ways. Several published reports have estimated the potential of the ACA to reduce the number of uninsured veterans by estimating how many uninsured veterans would be income eligible for either Medicaid or for tax credits through a Marketplace. The Urban Institute estimated that half of all uninsured veterans in 2010–2013 had incomes below 138% of the poverty level, but 400,000 of these had incomes below 100% of the poverty level, making them ineligible for financial assistance through a Health Insurance Marketplace for veterans living in states that did not expand Medicaid.12 A more recent report estimated that 25% of uninsured veterans would be eligible for Medicaid, even with half of the states not expanding their Medicaid programs, and 46% would be eligible for subsidized coverage through federal and state-based marketplace options.3 The 5.3 million veterans living in rural areas face specific health care concerns.13 Compared with urban veterans, rural veterans tend to have lower health status, including higher levels of hypertension, diabetes mellitus, hyperlipidemia, post-traumatic stress disorder, and depressive disorder.6 Rural veterans also face challenging social barriers to care including lack of transportation, substandard housing, low levels of education, high rates of unemployment and low rates of labor force participation.2 Rural veterans also are more likely to live further away from a VHA facility, and thus non-VHA coverage options may be particularly important for them.14 In this study, we examined changes in health insurance coverage during the first year of full implementation of the ACA (from 2013 to 2014) for non-elderly urban and rural veterans. Our analysis investigated if rural and urban veterans that lived in states that expanded Medicaid experienced more coverage gains than those that lived in states not expanding Medicaid coverage. We also examined if changes in coverage differed by income. METHODS We obtained data on veterans from the 2013 and 2014 American Community Survey (ACS), a nationally representative survey conducted by the US Census Bureau. In 2014, the Census Bureau reported a 96.7% response rate.15 The primary advantage of the ACS is its large sample size which allows for detailed analysis of relatively small subgroups such as veterans living in rural areas. In 2014, the public-use sample included 3.1 million person records. Ethical standards were used in this research. All data used in this study were publicly available secondary data with no individual identifiers. No human subjects were involved. Veterans We defined veterans as those who reported that they served “On active duty in the past, but not now.” We restricted our analysis to veterans age 18–64 that dwell in the community (i.e., are not institutionalized). Overall, our analytical sample size in the pooled 2013–2014 file was 199,398 persons. Rural and Urban Designation The lowest level of geography in the public-use ACS is the Public Use Microdata Area (PUMA), a Census geography that is state specific and contains approximately 100,000 people. In populated areas there are multiple PUMAs per county, but in sparsely populated regions, PUMAs consist of multiple counties. We defined rural as living in a PUMA in which at least 50% of the PUMA’s population lived in a rural Census Block, as defined by the Census Bureau. All other observations were considered to be urban. The rural/urban data were obtained from the Missouri Census Data Center.16 Median population density for the PUMAs we classified as urban was 1,613 people per square mile compared with 55 people per square mile for the rural PUMAs. Health Insurance Coverage We examined six types of health insurance coverage: (1) Employer-sponsored insurance (including Tri-Care); (2) Non-VHA public coverage (Medicare, Medicaid, and other public sources for those with low-incomes or a disability); (3) Coverage purchased directly from an insurance company which includes non-group coverage purchased on or off the exchanges; (4) Those obtaining care from the VHA, regardless of other coverage types they report; (5) Those who list no other source of coverage except VHA care; and (6) The uninsured (those not reporting any coverage type). We distinguished between veterans that had VHA care in any combination with other coverage and veterans that had only VHA coverage because changes in each group has different policy significance. In our results and tables we refer to services obtained from the VHA as VA coverage to be consistent with the ACS questionnaire wording (which uses “VA”). Other Covariates We included the following control variables: age (18–34, 35–44, 45–64); gender (male/female); race (non-Hispanic White, non-Hispanic Black, non-Hispanic Other Race, Hispanic); marital status (not married/married); educational attainment (high school graduate or less, some college or associates degree, Bachelor’s degree or more); employment (unemployed, part-time, full-time, or not in the labor force); functional difficulties arising from sensory, mobility, cognitive, or independent living limitations, service connected status (yes/no); and most recent era of military service (Vietnam or before, 1975–1990, Gulf Era, Since 9/11). We use the race/ethnicity classification included in the ACS. We created a measure of family income and determined poverty levels using Federal Poverty Guidelines (FPG) to create four income categories: 0–138% FPG, 139–249% FPG, 250–399% FPG, and ≥ 400% FPG. These groups correspond to the income thresholds for the new ACA Medicaid expansion population, the population eligible for premium tax credits and cost-sharing reductions in the Health Insurance Marketplaces, the population eligible for premium tax credits but not costs-sharing reductions, and those not eligible for any financial assistance, respectively. We also grouped our observations of veterans living in states that had expanded Medicaid by the start of 2014 versus those living in states that had not expanded Medicaid.17 Statistical Analysis We began by describing the characteristics of the urban and rural non-elderly veteran population in the baseline period (2013). Next, we examined unadjusted changes in coverage type between 2013 and 2014, overall and for urban and rural veterans separately. Finally, we estimated differences in the adjusted probability of being insured (any type of coverage) between 2013 and 2014. We stratified the sample by urban and rural status and computed these differences separately for each strata. Adjusted probabilities were computed by fitting logistic regressions that included year and controls for age, gender, race, marital status, poverty status, educational attainment, and employment. Two separate versions of the model were run: one that interacted year with Medicaid expansion status and another that interacted year with income category. We used t-tests to determine statistical significance at the p < 0.05 level. Coefficient estimates were converted to a probability scale for ease of interpretation.18 More details of our statistic model can be found in the Supplementary materials. Limitations Potential limitations to our study included the constraints imposed by the ACS on how we defined urban/rural status. Because the lowest level of geography was PUMAs and PUMA boundaries might encompass both urban and rural communities, our measure might have led to misclassification (we explore this issue more thoroughly in the Supplementary material). In addition, there was a change in the way veterans were identified in the ACS starting in 2013. As a result, we were only able to examine changes in coverage between 2013 and 2014. The lack of a substantial pre-period prevented us from determining if the changes we observed were simply a continuation of long-standing trends. However, data from other sources suggest stable rates going back at least to 2011.19 Furthermore, our results are in line with a growing body of literature which finds that the ACA is associated with increased coverage and access to care, independent of other factors.20 RESULTS Table I presents baseline characteristics of non-elderly veterans by urban and rural status. Table I. Characteristics of Community-Dwelling Non-Elderly Urban and Rural Veterans, 2013   Urban  Rural  Total  %  SE  %  SE  %  SE  2013 sample size  79,816  22,739  102,555  2013 population count  8,301,388  1,800,501  10,101,889  Region   Northeast  12.85  0.13  12.89  0.27  12.85  0.12   Midwest  19.31  0.18  29.79*  0.35  21.17  0.16   South  42.48  0.18  49.16*  0.38  43.67  0.17   West  25.37  0.16  8.17*  0.25  22.30  0.14  Age   18–34  16.45  0.15  12.10*  0.29  15.67  0.13   35–44  18.11  0.18  15.86*  0.30  17.71  0.15   45–64  65.45  0.20  72.04*  0.42  66.63  0.17  Gender   Male  87.02  0.13  89.08*  0.24  87.39  0.13   Female  12.98  0.13  10.92*  0.24  12.61  0.13  Race   White, Non-Hispanic  69.29  0.24  85.49*  0.30  72.18  0.21   Black, Non-Hispanic  16.51  0.15  8.47*  0.21  15.08  0.13   Other, Non-Hispanic  5.01  0.08  3.32*  0.15  4.71  0.08   Hispanic  9.18  0.16  2.73*  0.15  8.03  0.13  Marriage   Not married  39.27  0.23  35.86*  0.45  38.66  0.19   Married  60.73  0.23  64.14*  0.45  61.34  0.19  Family income   0–138% FPG  17.02  0.16  20.00*  0.36  17.55  0.14   139–249% FPG  15.58  0.14  18.98*  0.36  16.19  0.13   250–399% FPG  21.42  0.18  24.71*  0.32  22.01  0.16   ≥400% FPG  45.98  0.23  36.32*  0.40  44.26  0.20  Educational attainment   High School, GED, or less  28.88  0.20  39.59*  0.41  30.69  0.07   Some college  44.92  0.20  42.79*  0.45  44.53  0.18   Bachelor’s or more  26.32  0.19  17.62*  0.30  24.77  0.19  Employment   Unemployed  5.67  0.11  4.95*  0.20  5.54  0.10   Part-timer worker  4.56  0.08  4.41  0.18  4.53  0.07   Full-time worker  66.75  0.20  61.04*  0.39  65.73  0.17   Not in labor force  23.03  0.18  29.61*  0.39  24.20  0.16  Functional limitations   None  82.84  0.18  77.33*  0.38  81.86  0.16   Any  17.16  0.18  22.67*  0.38  18.14  0.16  Service connected   No  78.69  0.16  78.53  0.32  78.66  0.14   Yes  21.31  0.16  21.47  0.32  21.34  0.14  Most recent period of service   Vietnam era or earlier  16.56  0.14  12.58*  0.29  15.85  0.12   1975–1990  19.91  0.15  17.04*  0.35  19.40  0.13   Gulf War Era  37.04  0.21  37.11  0.40  37.05  0.18   Since 911  26.49  0.19  33.28*  0.35  27.70  0.17    Urban  Rural  Total  %  SE  %  SE  %  SE  2013 sample size  79,816  22,739  102,555  2013 population count  8,301,388  1,800,501  10,101,889  Region   Northeast  12.85  0.13  12.89  0.27  12.85  0.12   Midwest  19.31  0.18  29.79*  0.35  21.17  0.16   South  42.48  0.18  49.16*  0.38  43.67  0.17   West  25.37  0.16  8.17*  0.25  22.30  0.14  Age   18–34  16.45  0.15  12.10*  0.29  15.67  0.13   35–44  18.11  0.18  15.86*  0.30  17.71  0.15   45–64  65.45  0.20  72.04*  0.42  66.63  0.17  Gender   Male  87.02  0.13  89.08*  0.24  87.39  0.13   Female  12.98  0.13  10.92*  0.24  12.61  0.13  Race   White, Non-Hispanic  69.29  0.24  85.49*  0.30  72.18  0.21   Black, Non-Hispanic  16.51  0.15  8.47*  0.21  15.08  0.13   Other, Non-Hispanic  5.01  0.08  3.32*  0.15  4.71  0.08   Hispanic  9.18  0.16  2.73*  0.15  8.03  0.13  Marriage   Not married  39.27  0.23  35.86*  0.45  38.66  0.19   Married  60.73  0.23  64.14*  0.45  61.34  0.19  Family income   0–138% FPG  17.02  0.16  20.00*  0.36  17.55  0.14   139–249% FPG  15.58  0.14  18.98*  0.36  16.19  0.13   250–399% FPG  21.42  0.18  24.71*  0.32  22.01  0.16   ≥400% FPG  45.98  0.23  36.32*  0.40  44.26  0.20  Educational attainment   High School, GED, or less  28.88  0.20  39.59*  0.41  30.69  0.07   Some college  44.92  0.20  42.79*  0.45  44.53  0.18   Bachelor’s or more  26.32  0.19  17.62*  0.30  24.77  0.19  Employment   Unemployed  5.67  0.11  4.95*  0.20  5.54  0.10   Part-timer worker  4.56  0.08  4.41  0.18  4.53  0.07   Full-time worker  66.75  0.20  61.04*  0.39  65.73  0.17   Not in labor force  23.03  0.18  29.61*  0.39  24.20  0.16  Functional limitations   None  82.84  0.18  77.33*  0.38  81.86  0.16   Any  17.16  0.18  22.67*  0.38  18.14  0.16  Service connected   No  78.69  0.16  78.53  0.32  78.66  0.14   Yes  21.31  0.16  21.47  0.32  21.34  0.14  Most recent period of service   Vietnam era or earlier  16.56  0.14  12.58*  0.29  15.85  0.12   1975–1990  19.91  0.15  17.04*  0.35  19.40  0.13   Gulf War Era  37.04  0.21  37.11  0.40  37.05  0.18   Since 911  26.49  0.19  33.28*  0.35  27.70  0.17  Source: 2013 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census block, as defined by the Census Bureau. SE are standard errors estimated using successive difference replication. *Indicates a statistically significant difference between urban and rural at the p< 0.05 level. Table I. Characteristics of Community-Dwelling Non-Elderly Urban and Rural Veterans, 2013   Urban  Rural  Total  %  SE  %  SE  %  SE  2013 sample size  79,816  22,739  102,555  2013 population count  8,301,388  1,800,501  10,101,889  Region   Northeast  12.85  0.13  12.89  0.27  12.85  0.12   Midwest  19.31  0.18  29.79*  0.35  21.17  0.16   South  42.48  0.18  49.16*  0.38  43.67  0.17   West  25.37  0.16  8.17*  0.25  22.30  0.14  Age   18–34  16.45  0.15  12.10*  0.29  15.67  0.13   35–44  18.11  0.18  15.86*  0.30  17.71  0.15   45–64  65.45  0.20  72.04*  0.42  66.63  0.17  Gender   Male  87.02  0.13  89.08*  0.24  87.39  0.13   Female  12.98  0.13  10.92*  0.24  12.61  0.13  Race   White, Non-Hispanic  69.29  0.24  85.49*  0.30  72.18  0.21   Black, Non-Hispanic  16.51  0.15  8.47*  0.21  15.08  0.13   Other, Non-Hispanic  5.01  0.08  3.32*  0.15  4.71  0.08   Hispanic  9.18  0.16  2.73*  0.15  8.03  0.13  Marriage   Not married  39.27  0.23  35.86*  0.45  38.66  0.19   Married  60.73  0.23  64.14*  0.45  61.34  0.19  Family income   0–138% FPG  17.02  0.16  20.00*  0.36  17.55  0.14   139–249% FPG  15.58  0.14  18.98*  0.36  16.19  0.13   250–399% FPG  21.42  0.18  24.71*  0.32  22.01  0.16   ≥400% FPG  45.98  0.23  36.32*  0.40  44.26  0.20  Educational attainment   High School, GED, or less  28.88  0.20  39.59*  0.41  30.69  0.07   Some college  44.92  0.20  42.79*  0.45  44.53  0.18   Bachelor’s or more  26.32  0.19  17.62*  0.30  24.77  0.19  Employment   Unemployed  5.67  0.11  4.95*  0.20  5.54  0.10   Part-timer worker  4.56  0.08  4.41  0.18  4.53  0.07   Full-time worker  66.75  0.20  61.04*  0.39  65.73  0.17   Not in labor force  23.03  0.18  29.61*  0.39  24.20  0.16  Functional limitations   None  82.84  0.18  77.33*  0.38  81.86  0.16   Any  17.16  0.18  22.67*  0.38  18.14  0.16  Service connected   No  78.69  0.16  78.53  0.32  78.66  0.14   Yes  21.31  0.16  21.47  0.32  21.34  0.14  Most recent period of service   Vietnam era or earlier  16.56  0.14  12.58*  0.29  15.85  0.12   1975–1990  19.91  0.15  17.04*  0.35  19.40  0.13   Gulf War Era  37.04  0.21  37.11  0.40  37.05  0.18   Since 911  26.49  0.19  33.28*  0.35  27.70  0.17    Urban  Rural  Total  %  SE  %  SE  %  SE  2013 sample size  79,816  22,739  102,555  2013 population count  8,301,388  1,800,501  10,101,889  Region   Northeast  12.85  0.13  12.89  0.27  12.85  0.12   Midwest  19.31  0.18  29.79*  0.35  21.17  0.16   South  42.48  0.18  49.16*  0.38  43.67  0.17   West  25.37  0.16  8.17*  0.25  22.30  0.14  Age   18–34  16.45  0.15  12.10*  0.29  15.67  0.13   35–44  18.11  0.18  15.86*  0.30  17.71  0.15   45–64  65.45  0.20  72.04*  0.42  66.63  0.17  Gender   Male  87.02  0.13  89.08*  0.24  87.39  0.13   Female  12.98  0.13  10.92*  0.24  12.61  0.13  Race   White, Non-Hispanic  69.29  0.24  85.49*  0.30  72.18  0.21   Black, Non-Hispanic  16.51  0.15  8.47*  0.21  15.08  0.13   Other, Non-Hispanic  5.01  0.08  3.32*  0.15  4.71  0.08   Hispanic  9.18  0.16  2.73*  0.15  8.03  0.13  Marriage   Not married  39.27  0.23  35.86*  0.45  38.66  0.19   Married  60.73  0.23  64.14*  0.45  61.34  0.19  Family income   0–138% FPG  17.02  0.16  20.00*  0.36  17.55  0.14   139–249% FPG  15.58  0.14  18.98*  0.36  16.19  0.13   250–399% FPG  21.42  0.18  24.71*  0.32  22.01  0.16   ≥400% FPG  45.98  0.23  36.32*  0.40  44.26  0.20  Educational attainment   High School, GED, or less  28.88  0.20  39.59*  0.41  30.69  0.07   Some college  44.92  0.20  42.79*  0.45  44.53  0.18   Bachelor’s or more  26.32  0.19  17.62*  0.30  24.77  0.19  Employment   Unemployed  5.67  0.11  4.95*  0.20  5.54  0.10   Part-timer worker  4.56  0.08  4.41  0.18  4.53  0.07   Full-time worker  66.75  0.20  61.04*  0.39  65.73  0.17   Not in labor force  23.03  0.18  29.61*  0.39  24.20  0.16  Functional limitations   None  82.84  0.18  77.33*  0.38  81.86  0.16   Any  17.16  0.18  22.67*  0.38  18.14  0.16  Service connected   No  78.69  0.16  78.53  0.32  78.66  0.14   Yes  21.31  0.16  21.47  0.32  21.34  0.14  Most recent period of service   Vietnam era or earlier  16.56  0.14  12.58*  0.29  15.85  0.12   1975–1990  19.91  0.15  17.04*  0.35  19.40  0.13   Gulf War Era  37.04  0.21  37.11  0.40  37.05  0.18   Since 911  26.49  0.19  33.28*  0.35  27.70  0.17  Source: 2013 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census block, as defined by the Census Bureau. SE are standard errors estimated using successive difference replication. *Indicates a statistically significant difference between urban and rural at the p< 0.05 level. Overall, we estimated 10.1 million U.S. non-elderly veterans in 2013; 82% lived in predominantly urban PUMAs (8.3 million), and the remaining 18% (1.8 million) lived in predominately rural PUMAs. Most veterans lived in the South (43.6%), and rural veterans were more likely to be Southerners than their urban counterparts (P < 0.05). On every marker of economic well-being, rural veterans fared worse than urban veterans. They had a statistically significant higher chance of having incomes being below 138% of FPG (20.0% versus 17.0%), of being out of the labor force (29.1% versus 23.0%), and of having no more than a high school education (39.6% versus 28.8%). Rural veterans were also more likely to experience at least one functional limitation, but they had the same chance of having a service connected disability compared to their urban counterparts. Table II presents unadjusted estimates of health insurance type by year, overall and separately for urban and rural veterans. Non-elderly veterans, regardless of urban/rural status, experienced a significant 2.2 percentage point reduction in uninsurance during the first year of the ACA (from 2013 to 2014). This translated into a reduction in the number of uninsured Veterans by approximately 230,000. Overall, coverage gains were driven by increases in employer-sponsored coverage (0.8 percentage points) and non-VA public coverage (primarily Medicaid) (0.8 percentage points). There was also a significant increase in the number of veterans accessing VHA care, but there was no change in the percent of veterans who only had VHA care and no other coverage type. Table II. Changes in Coverage Type for Community-Dwelling Non-Elderly Urban and Rural Vets, 2013–2014   2013  2014  Difference  %  SE  %  SE    Total sample size  102,555    96,843      Total population size  10,101,889    9,715,012      Total   Employer/tri-care coverage  70.69  0.21  71.47  0.17  0.79*   Medicaid/Medicare/other public  11.73  0.11  12.57  0.14  0.83*   Directly purchased coverage  7.70  0.11  7.97  0.11  0.28   VA coverage (Any)  27.22  0.20  28.17  0.18  0.95*   VA coverage (Alone)  8.22  0.12  8.22  0.11  0.00   Uninsured  9.31  0.12  7.11  0.11  −2.2*  Urban   Employer/tri-care coverage  71.69  0.22  72.62  0.19  0.92*   Medicaid/Medicare/other public  11.01  0.12  11.76  0.15  0.75*   Directly purchased coverage  7.68  0.12  7.85  0.12  0.17   VA coverage (any)  26.68  0.22  27.68  0.22  0.99*   VA coverage (alone)  8.04  0.14  8.11  0.12  0.07   Uninsured  9.01  0.13  6.76  0.12  −2.25*  Rural   Employer/tri-care coverage  66.06  0.40  66.18  0.44  0.12   Medicaid/Medicare/other public  15.08  0.33  16.31  0.33  1.23*   Directly purchased coverage  7.77  0.23  8.53  0.24  0.76*   VA coverage (any)  29.72  0.36  30.48  0.39  0.76   VA coverage (alone)  9.04  0.23  8.72  0.25  −0.33   Uninsured  10.73  0.29  8.74  0.25  −1.99*    2013  2014  Difference  %  SE  %  SE    Total sample size  102,555    96,843      Total population size  10,101,889    9,715,012      Total   Employer/tri-care coverage  70.69  0.21  71.47  0.17  0.79*   Medicaid/Medicare/other public  11.73  0.11  12.57  0.14  0.83*   Directly purchased coverage  7.70  0.11  7.97  0.11  0.28   VA coverage (Any)  27.22  0.20  28.17  0.18  0.95*   VA coverage (Alone)  8.22  0.12  8.22  0.11  0.00   Uninsured  9.31  0.12  7.11  0.11  −2.2*  Urban   Employer/tri-care coverage  71.69  0.22  72.62  0.19  0.92*   Medicaid/Medicare/other public  11.01  0.12  11.76  0.15  0.75*   Directly purchased coverage  7.68  0.12  7.85  0.12  0.17   VA coverage (any)  26.68  0.22  27.68  0.22  0.99*   VA coverage (alone)  8.04  0.14  8.11  0.12  0.07   Uninsured  9.01  0.13  6.76  0.12  −2.25*  Rural   Employer/tri-care coverage  66.06  0.40  66.18  0.44  0.12   Medicaid/Medicare/other public  15.08  0.33  16.31  0.33  1.23*   Directly purchased coverage  7.77  0.23  8.53  0.24  0.76*   VA coverage (any)  29.72  0.36  30.48  0.39  0.76   VA coverage (alone)  9.04  0.23  8.72  0.25  −0.33   Uninsured  10.73  0.29  8.74  0.25  −1.99*  Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census block, as defined by the Census Bureau. Coverage types are not mutually exclusive, see text for details. SE are standard errors estimated using successive difference replication. *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. Note that none of the changes over time were statistically different in rural versus urban areas. Table II. Changes in Coverage Type for Community-Dwelling Non-Elderly Urban and Rural Vets, 2013–2014   2013  2014  Difference  %  SE  %  SE    Total sample size  102,555    96,843      Total population size  10,101,889    9,715,012      Total   Employer/tri-care coverage  70.69  0.21  71.47  0.17  0.79*   Medicaid/Medicare/other public  11.73  0.11  12.57  0.14  0.83*   Directly purchased coverage  7.70  0.11  7.97  0.11  0.28   VA coverage (Any)  27.22  0.20  28.17  0.18  0.95*   VA coverage (Alone)  8.22  0.12  8.22  0.11  0.00   Uninsured  9.31  0.12  7.11  0.11  −2.2*  Urban   Employer/tri-care coverage  71.69  0.22  72.62  0.19  0.92*   Medicaid/Medicare/other public  11.01  0.12  11.76  0.15  0.75*   Directly purchased coverage  7.68  0.12  7.85  0.12  0.17   VA coverage (any)  26.68  0.22  27.68  0.22  0.99*   VA coverage (alone)  8.04  0.14  8.11  0.12  0.07   Uninsured  9.01  0.13  6.76  0.12  −2.25*  Rural   Employer/tri-care coverage  66.06  0.40  66.18  0.44  0.12   Medicaid/Medicare/other public  15.08  0.33  16.31  0.33  1.23*   Directly purchased coverage  7.77  0.23  8.53  0.24  0.76*   VA coverage (any)  29.72  0.36  30.48  0.39  0.76   VA coverage (alone)  9.04  0.23  8.72  0.25  −0.33   Uninsured  10.73  0.29  8.74  0.25  −1.99*    2013  2014  Difference  %  SE  %  SE    Total sample size  102,555    96,843      Total population size  10,101,889    9,715,012      Total   Employer/tri-care coverage  70.69  0.21  71.47  0.17  0.79*   Medicaid/Medicare/other public  11.73  0.11  12.57  0.14  0.83*   Directly purchased coverage  7.70  0.11  7.97  0.11  0.28   VA coverage (Any)  27.22  0.20  28.17  0.18  0.95*   VA coverage (Alone)  8.22  0.12  8.22  0.11  0.00   Uninsured  9.31  0.12  7.11  0.11  −2.2*  Urban   Employer/tri-care coverage  71.69  0.22  72.62  0.19  0.92*   Medicaid/Medicare/other public  11.01  0.12  11.76  0.15  0.75*   Directly purchased coverage  7.68  0.12  7.85  0.12  0.17   VA coverage (any)  26.68  0.22  27.68  0.22  0.99*   VA coverage (alone)  8.04  0.14  8.11  0.12  0.07   Uninsured  9.01  0.13  6.76  0.12  −2.25*  Rural   Employer/tri-care coverage  66.06  0.40  66.18  0.44  0.12   Medicaid/Medicare/other public  15.08  0.33  16.31  0.33  1.23*   Directly purchased coverage  7.77  0.23  8.53  0.24  0.76*   VA coverage (any)  29.72  0.36  30.48  0.39  0.76   VA coverage (alone)  9.04  0.23  8.72  0.25  −0.33   Uninsured  10.73  0.29  8.74  0.25  −1.99*  Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census block, as defined by the Census Bureau. Coverage types are not mutually exclusive, see text for details. SE are standard errors estimated using successive difference replication. *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. Note that none of the changes over time were statistically different in rural versus urban areas. The decline in the uninsured was similar for both urban and rural populations. However, the specific types of coverage driving the reduction in uninsured differed between urban and rural status. Veterans in urban areas experienced a significant increase in employer-sponsored insurance of 1 percentage point and significant increase in non-VA public coverage of 0.75 percentage points. Rural veterans did not experience a significant change in employer-sponsored insurance, but they did experience a significant gain in non-VA public coverage (1.23 percentage points). Rural veterans also experienced a statistically significant increase in direct purchase coverage while the change in direct purchase coverage for urban veterans was small and not significant. In Figures 1 and 2, we present adjusted differences in the percent of veterans that were insured between 2013 and 2014. Complete regression results are provided in the Supplementary material. Figure 1 describes changes in the percent insured by urban/rural status, for veterans living in Medicaid expansion states versus those living in non-expansion states. Overall, veterans in Medicaid expansion states experienced a significantly larger increase in insurance compared with veterans living in non-expansion states. Rural veterans in Medicaid expansion states had significantly larger increases compared to rural veterans in non-expansion states. For rural veterans in Medicaid expansion states, the increase in insurance was 3.5 percentage points, compared with 1.2 percentage points in non-expansion states. While all urban veterans experienced a significant increase in insurance, the relative increase in Medicaid expansion versus non- expansion states was considerably smaller than it was for rural veterans, and it was not significant. Figure 1. View largeDownload slide Differences in the Adjusted Percent Insured Between 2014 and 2013, by Urban/Rural and Expansion Status. Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census Block, as defined by the Census Bureau. Results represent average marginal effects from 3 separate logistic regressions (see text for details). *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. †indicates that 2014–2013 difference is significantly different in Medicaid expansion versus non-expansion states (P < 0.05). Please see the Supplementary material for complete regression results. Figure 1. View largeDownload slide Differences in the Adjusted Percent Insured Between 2014 and 2013, by Urban/Rural and Expansion Status. Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census Block, as defined by the Census Bureau. Results represent average marginal effects from 3 separate logistic regressions (see text for details). *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. †indicates that 2014–2013 difference is significantly different in Medicaid expansion versus non-expansion states (P < 0.05). Please see the Supplementary material for complete regression results. FIGURE 2. View largeDownload slide Differences in the Adjusted Percent Insured Between 2014 and 2013, by Urban/Rural and Income. Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census Block, as defined by the Census Bureau. Results represent average marginal effects from three separate logistic regressions (see text for details). *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. †Indicates that 2014–2013 difference is significantly different in the given income category compared to the lowest-income group (P < 0.05). Please see the Supplementary material for complete regression results. FIGURE 2. View largeDownload slide Differences in the Adjusted Percent Insured Between 2014 and 2013, by Urban/Rural and Income. Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census Block, as defined by the Census Bureau. Results represent average marginal effects from three separate logistic regressions (see text for details). *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. †Indicates that 2014–2013 difference is significantly different in the given income category compared to the lowest-income group (P < 0.05). Please see the Supplementary material for complete regression results. In Figure 2, we report adjusted differences by urban/rural and family income categories. Overall, those in the lowest-income group experienced the largest percentage point increase in insurance. For rural veterans, the lowest-income group had the largest increases followed by those in the 251–399% FPG group. The change for the 251–399% FPG group was not statistically different from the change in the 0–138% FPG group. We conducted additional subgroup analyses and explored how sensitive our results were to various modeling approaches (see Supplementary material). For example, because the ACA targeted those without access to employer-sponsored insurance, we examined regression adjusted changes in insurance for low-income urban and rural veterans by employment status. Non-working urban veterans experienced a 6.8 percentage point increase in insurance and non-working rural veterans experienced an 8.4 point increase. Differences for working veterans were smaller in both urban and rural communities. We also explored how sensitive our main results were to alternative modeling assumptions. We found that removing the states that were early adopters of the Medicaid expansion (expanding prior to January 1, 2014) had little impact on our Medicaid expansion results. We also obtained similar results when our Medicaid expansion regressions included state fixed effects. Finally, we explored how our results changed as we lowered the threshold at which we considered a veteran to reside in a predominantly rural community from 50% to 25%. Lowering the threshold had no effect on estimates of declines in uninsurance by urban and rural status, but as we reduced the threshold we found that changes to specific coverage types tended to appear similar across rural and urban areas. This did not surprise us, as lowering the threshold had the effect of increasing the “urban-ness” of those PUMAs we initially classified as predominantly rural. CONCLUSION There was a significant increase in insurance for U.S. veterans during the first year of the ACA. Our results suggest a substantial 24% relative decline in uninsurance from 2013 to 2014 – from a 9.3% uninsured rate to 7.1%. The changes in insurance status we observed were similar for both urban and rural veterans, but coverage gains by rural/urban status were driven by different coverage pathways. In urban areas, the increase in coverage was driven in large part by gains in employer-sponsored insurance that was likely the result of an improving economy and not the direct effect of the ACA. In rural areas, the increase in coverage was largely due to gains in Medicaid and individual market coverage. Residence in a Medicaid expansion state was particularly influential for rural veterans – the decline in the uninsured rate was three times larger in Medicaid expansion states versus non-expansion states. Our analysis suggests that Medicaid expansion was associated with a significant increase in health insurance coverage rates for rural veterans; we did not find a similar association for urban veterans. Rural veterans had neither gains nor losses in employer-sponsored insurance. Relatedly, we found that rural veterans were more likely to be out of the labor force compared to urban veterans and thus Medicaid may have played an important role in buffering the negative impact of a weak labor market. Two-thirds of non-elderly U.S. veterans were either Gulf War veterans or in service since 9/11, between the ages of 45 and 64, and working full time. Not surprisingly, most veterans (71.5% in 2014) obtained their health insurance through their or a spouse’s place of employment, higher than the rate for the non-elderly with employer-sponsored insurance for the general U.S. population (56%).21 Yet for low-income veterans including those without jobs or employer-sponsored insurance, the ACA has provided new pathways for obtaining health insurance coverage which have supplements services and financing provided through the VHA. We found that the Medicaid expansion was a particular important source of coverage for rural veterans. A concern at this point is for the many (43.7%) veterans living in the south where 10 out of 19 states (as of 2016) that have not expanded Medicaid22 are located.23 U.S. veterans rely on both public and private health insurance coverage in addition to VHA coverage. As policymakers and VA administrators continue to seek and experiment with new policy alternatives for increasing access to health services for veterans, the new coverage options created by the ACA will be an important tool. Perhaps more importantly, states that have not yet expanded Medicaid should consider the impact that expansion would have on increased coverage and access to care for low-income veterans and particularly those residing in rural areas. Supplementary Material Supplementary material is available at Military Medicine online. Acknowledgement This study was funded in part under a grant from the Robert Wood Johnson Foundation to the State Health Access Data Assistance Center at the University of Minnesota School of Public Health. Funding This study was funded in part under a grant from the Robert Wood Johnson Foundation to the State Health Access Data Assistance Center at the University of Minnesota School of Public Health. References 1 VA health care 2016: Developing solutions for the next generation of veteran care. International Quality & Productivity Center. Available at http://www.veteransaffairshealthcare.com/; accessed March 5, 2016. 2 Choski DA, Sommers B: Universal health coverage for US veterans: a goal within reach. Lancet  2014; 385( 9984): 2320– 1. Google Scholar PubMed  3 Haley J, Kenney GM; The Urban Institute. Uninsured veterans and family members: state and national estimates of expanded Medicaid eligibility under the ACA. Available at http://research.urban.org/UploadedPDF/412775-Uninsured-Veterans-and-Family-Members.pdf. 2013; accessed March 5, 2016. 4 Federal Register (US). Annual update of the HHS Federal Poverty Guidelines. 2014;70 FR(3593):2593–4. Available at https://federalregister.gov/a/2014-01303. 2014; accessed August 24, 2016. 5 Henry J Kaiser Family Foundation. Status of state action on the Medicaid expansion decision. Available at http://kff.org/health-reform/state-indicator/state-activity-around-expanding-medicaid-under-the-affordable-care-act/. 2016; accessed March 5, 2016. 6 Sommers BD, Gunja M, Finegold K, et al.  : Changes in self-reported Insurance coverage, access to care, and health under the affordable care act. JAMA.  2015; 314( 4): 366– 74. Google Scholar CrossRef Search ADS PubMed  7 Sommers BD, Maylone B, Nguyen KH, et al.  : The impact of state policies on ACA applications and enrollment among low-income adults in Arkansas, Kentucky, and Texas. Health Affairs  2015; 34( 6): 1010– 8. Google Scholar CrossRef Search ADS PubMed  8 Chen J, Vargas-Bustamante A, Mortensen K, et al.  : Racial and ethnic disparities in health care access and utilization under the Affordable Care Act. Med Care  2016; 54( 2): 140– 6. Google Scholar CrossRef Search ADS PubMed  9 Shartzer A, Long SK, Karpman M, et al.  : Quicktake: Insurance coverage gains cross economic, social, and geographic boundaries. Health Reform Monitoring Survey. Available at http://hrms.urban.org/quicktakes/Insurance-Coverage-Gains-Cross-Economic-Social-and-Geographic-Boundaries.html. 2015; accessed March 5, 2016. 10 Drake C, Abraham JM, McCullough JS: Rural enrollment in the federally facilitated marketplace. J Rural Health  2016; 32( 3): 332– 9. Google Scholar CrossRef Search ADS PubMed  11 Gooptu A, Moriya AS, Simon KI, et al.  : Medicaid expansion did not result in significant employment changes or job reductions in 2014. Health Aff  2016; 35( 1): 111– 8. Google Scholar CrossRef Search ADS   12 Haley J, Kenney GM; The Urban Institute. Uninsured veterans and family members: who are they and where do they live? Available at http://www.urban.org/research/publication/uninsured-veterans-and-family-members-who-are-they-and-where-do-they-live. 2012; accessed March 5, 2016. 13 Department of Veterans Affairs. About rural veterans. Available at http://www.ruralhealth.va.gov/about/rural-veterans.asp; accessed March 5, 2016. 14 Buzza C, Ono SS, Turvey C, et al.  : Distance is relative: unpacking a principal barrier in rural healthcare. J Gen Intern Med  2011; 26( Suppl 2): 648– 54. doi:10.1007/s11606-011-1762-1; accessed March 5, 2016. Google Scholar CrossRef Search ADS PubMed  15 The United States Census Bureau. American Community Survey response rates. Available at http://www.census.gov/acs/www/methodology/sample-size-and-data-quality/response-rates/; accessed March 5, 2016. 16 Missouri Census Data Center. MABLE/Geocorr12, Version 12: Geographic correspondence engine. Available at http://mcdc.missouri.edu/websas/geocorr12.html; accessed February 2016. 17 Henry J Kaiser Family Foundation. Medicaid eligibility for adults as of January 1, 2014. Available at http://kff.org/medicaid/fact-sheet/medicaid-eligibility-for-adults-as-of-january-1-2014/. 2013 accessed March 5, 2016. 18 Karaca-Mandic P, Norton E, Dowd B: Interaction terms in nonlinear models. Health Serv Res  2012; 47( 1.1): 255– 74. Google Scholar CrossRef Search ADS PubMed  19 Haley J, Kenney G; The Urban Institute. The Veterans Day, early evidence that the ACA is serving those who have served. Available at http://www.urban.org/urban-wire/veterans-day-early-evidence-aca-serving-those-who-have-served. 2015; accessed March 5, 2016. 20 French MT, Homer J, Gumus G, et al.  : Key provisions of the Patient Protection and Affordable Care Act (ACA): A systemic review and presentation of early research findings. Health Serv Res  2016; 51( 5): 1735– 71. doi:10.1111/1475-6773.12511; accessed March 5, 2016. Google Scholar CrossRef Search ADS PubMed  21 Henry J Kaiser Family Foundation. Health insurance coverage of nonelderly 0–64. Available at http://kff.org/other/state-indicator/nonelderly-0-64/?currentTimeframe=0&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D; accessed August 30, 2016. 22 Families USA. A 50-state look at Medicaid expansion. Available at http://familiesusa.org/product/50-state-look-medicaid-expansion. 2016; accessed August 30, 2016. 23 U.S. Census Bureau. Census Bureau regions and divisions with state FIPS codes. Available at https://www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf; accessed August 30, 2016. Author notes The views expressed are those of the authors and do not necessarily represent the views of the Robert Wood Johnson Foundation. © Association of Military Surgeons of the United States 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Military Medicine Oxford University Press

Recent Changes in Health Insurance Coverage for Urban and Rural Veterans: Evidence from the First Year of the Affordable Care Act

Loading next page...
 
/lp/ou_press/recent-changes-in-health-insurance-coverage-for-urban-and-rural-4MVndE0OWz
Publisher
Association of Military Surgeons of the United States
Copyright
© Association of Military Surgeons of the United States 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0026-4075
eISSN
1930-613X
D.O.I.
10.1093/milmed/usy053
Publisher site
See Article on Publisher Site

Abstract

Abstract Introduction Prior to the Affordable Care Act, as many as 1.3 million veterans lacked health insurance. With the passage of the Affordable Care Act, veterans now have new pathways to coverage through Medicaid expansion in those states that chose to expand Medicaid and through private coverage options offered through the Health Insurance Marketplace. We examined the impact of the ACA on health insurance coverage for veterans in expansion and non-expansion states and for urban and rural veterans. Methods We examined changes in veterans’ health insurance coverage following the first year of the ACA, focusing on whether they lived in an urban or rural area and whether they live in a Medicaid expansion state. We used data on approximately 200,000 non-elderly community-dwelling veterans, obtained from the 2013–2014 American Community Survey and estimated differences in the adjusted probability of being uninsured between 2013 and 2014 for both urban and rural areas. Adjusted probabilities were computed by fitting logistic regressions controlling for age, gender, race, marital status, poverty status, education, and employment. Results There were an estimated 10.1 million U.S. non-elderly veterans in 2013; 82% lived in predominantly urban areas (8.3 million), and the remaining 18% (1.8 million) lived in predominately rural areas. Most veterans lived in the South (43.6%), and rural veterans were more likely to be Southerners than their urban counterparts. On every marker of economic well-being, rural veterans fared worse than urban veterans. They had a statistically significant higher chance of having incomes below 138% of FPG (20.0% versus 17.0%), of being out of the labor force (29.1% versus 23.0%), and of having no more than a high school education (39.6% versus 28.8%). Rural veterans were also more likely to experience at least one functional limitation. Overall, veterans in Medicaid expansion states experienced a significantly larger increase in insurance compared to veterans living in non-expansion states. For rural veterans in Medicaid expansion states, the increase in insurance was 3.5 percentage points, compared with 1.2 percentage points in non-expansion states. Conclusion Our analysis found a substantial 24% relative decline in the rate of uninsurance for U.S. Veterans, from 9.3 to 7.1% between 2013 and 2014. We found that coverage gains in rural areas were due to gains in Medicaid and individual market coverage. Residence in a Medicaid expansion state was particularly influential for rural veterans – the increase in the insured rate was three times larger in Medicaid expansion states versus non-expansion states. The ACA has had a positive and significant impact on the ability of U.S. Veterans to obtain health insurance coverage specifically for low-income veterans living in rural areas. The poverty rate among Veterans is rising and is particularly an issue for the more recent Gulf War veterans. Providing affordable and accessible health insurance options is part of our commitment to those who have served our country. Our analysis also presents yet another reason for the 17 non-expansion states to consider a Medicaid expansion. INTRODUCTION There are more veterans living in the USA today than at any other time since WWII.1 While the Department of Veterans’ Affairs (VA) plays an important role in providing health care services to veterans through the Veterans Health Administration (VHA), 55% of veterans rely on care obtained through employer-sponsored health insurance or public programs like Medicare and Medicaid.2 In addition, approximately 1.3 million non-elderly veterans lacked comprehensive health insurance and did not access VHA health services in 2010.3 Given that many veterans are covered outside of the VHA while others lack any coverage at all, recent health reform initiatives have the potential to substantially affect veterans’ access to health care. The passage of the Affordable Care Act (ACA) in 2010 created new coverage options for veterans. Eligible veterans may now obtain health insurance coverage through newly formed Health Insurance Marketplaces with tax credits for those that qualify. The ACA also allowed states to expand their Medicaid program to non-elderly adults with incomes below 138% of the poverty level (or $16,100 of annual income for a single adult in 2014).4 As of early 2016, 31 states and the District of Columbia expanded Medicaid under the Affordable Care Act, 2 states are considering expansion, and 17 states have no plans to expand.5 Although the full implementation of the ACA is only in its third year (as of 2016), recent research suggests that, in the general population, it is having substantial impacts on health insurance coverage, access to care, and self-reported health.6–9 Recent estimates from the National Health Interview Survey (NHIS) suggest that between 2013 and the first three quarters of 2015, the number of uninsured persons declined by 16 million.9 Reductions in the uninsured have been the largest for people with lower incomes, Hispanics, and residents of states that opted to expand Medicaid. There is little evidence to date regarding changes in coverage for rural versus urban populations, but one recent study suggests that individuals in rural areas were less likely to sign up for coverage in the Health Insurance Marketplace compared to people in urban areas.10 Reductions in the uninsured have been accompanied by increases in the number of non- elderly adults that report having a regular physician, a decrease in the fraction of non-elderly adults that needed but could not afford care, and a decrease in the number that report being in fair or poor health.6 To date, there is limited evidence of substantial declines in labor force participation,11 and the rate of employer-sponsored insurance has also remained steady, suggesting that consumers have not substituted private coverage for new public coverage options (i.e., “crowd-out”).9 Given their distinct demographic profile, health needs, and the dedicated health systems serving them, veterans are likely to be affected by the ACA in unique ways. Several published reports have estimated the potential of the ACA to reduce the number of uninsured veterans by estimating how many uninsured veterans would be income eligible for either Medicaid or for tax credits through a Marketplace. The Urban Institute estimated that half of all uninsured veterans in 2010–2013 had incomes below 138% of the poverty level, but 400,000 of these had incomes below 100% of the poverty level, making them ineligible for financial assistance through a Health Insurance Marketplace for veterans living in states that did not expand Medicaid.12 A more recent report estimated that 25% of uninsured veterans would be eligible for Medicaid, even with half of the states not expanding their Medicaid programs, and 46% would be eligible for subsidized coverage through federal and state-based marketplace options.3 The 5.3 million veterans living in rural areas face specific health care concerns.13 Compared with urban veterans, rural veterans tend to have lower health status, including higher levels of hypertension, diabetes mellitus, hyperlipidemia, post-traumatic stress disorder, and depressive disorder.6 Rural veterans also face challenging social barriers to care including lack of transportation, substandard housing, low levels of education, high rates of unemployment and low rates of labor force participation.2 Rural veterans also are more likely to live further away from a VHA facility, and thus non-VHA coverage options may be particularly important for them.14 In this study, we examined changes in health insurance coverage during the first year of full implementation of the ACA (from 2013 to 2014) for non-elderly urban and rural veterans. Our analysis investigated if rural and urban veterans that lived in states that expanded Medicaid experienced more coverage gains than those that lived in states not expanding Medicaid coverage. We also examined if changes in coverage differed by income. METHODS We obtained data on veterans from the 2013 and 2014 American Community Survey (ACS), a nationally representative survey conducted by the US Census Bureau. In 2014, the Census Bureau reported a 96.7% response rate.15 The primary advantage of the ACS is its large sample size which allows for detailed analysis of relatively small subgroups such as veterans living in rural areas. In 2014, the public-use sample included 3.1 million person records. Ethical standards were used in this research. All data used in this study were publicly available secondary data with no individual identifiers. No human subjects were involved. Veterans We defined veterans as those who reported that they served “On active duty in the past, but not now.” We restricted our analysis to veterans age 18–64 that dwell in the community (i.e., are not institutionalized). Overall, our analytical sample size in the pooled 2013–2014 file was 199,398 persons. Rural and Urban Designation The lowest level of geography in the public-use ACS is the Public Use Microdata Area (PUMA), a Census geography that is state specific and contains approximately 100,000 people. In populated areas there are multiple PUMAs per county, but in sparsely populated regions, PUMAs consist of multiple counties. We defined rural as living in a PUMA in which at least 50% of the PUMA’s population lived in a rural Census Block, as defined by the Census Bureau. All other observations were considered to be urban. The rural/urban data were obtained from the Missouri Census Data Center.16 Median population density for the PUMAs we classified as urban was 1,613 people per square mile compared with 55 people per square mile for the rural PUMAs. Health Insurance Coverage We examined six types of health insurance coverage: (1) Employer-sponsored insurance (including Tri-Care); (2) Non-VHA public coverage (Medicare, Medicaid, and other public sources for those with low-incomes or a disability); (3) Coverage purchased directly from an insurance company which includes non-group coverage purchased on or off the exchanges; (4) Those obtaining care from the VHA, regardless of other coverage types they report; (5) Those who list no other source of coverage except VHA care; and (6) The uninsured (those not reporting any coverage type). We distinguished between veterans that had VHA care in any combination with other coverage and veterans that had only VHA coverage because changes in each group has different policy significance. In our results and tables we refer to services obtained from the VHA as VA coverage to be consistent with the ACS questionnaire wording (which uses “VA”). Other Covariates We included the following control variables: age (18–34, 35–44, 45–64); gender (male/female); race (non-Hispanic White, non-Hispanic Black, non-Hispanic Other Race, Hispanic); marital status (not married/married); educational attainment (high school graduate or less, some college or associates degree, Bachelor’s degree or more); employment (unemployed, part-time, full-time, or not in the labor force); functional difficulties arising from sensory, mobility, cognitive, or independent living limitations, service connected status (yes/no); and most recent era of military service (Vietnam or before, 1975–1990, Gulf Era, Since 9/11). We use the race/ethnicity classification included in the ACS. We created a measure of family income and determined poverty levels using Federal Poverty Guidelines (FPG) to create four income categories: 0–138% FPG, 139–249% FPG, 250–399% FPG, and ≥ 400% FPG. These groups correspond to the income thresholds for the new ACA Medicaid expansion population, the population eligible for premium tax credits and cost-sharing reductions in the Health Insurance Marketplaces, the population eligible for premium tax credits but not costs-sharing reductions, and those not eligible for any financial assistance, respectively. We also grouped our observations of veterans living in states that had expanded Medicaid by the start of 2014 versus those living in states that had not expanded Medicaid.17 Statistical Analysis We began by describing the characteristics of the urban and rural non-elderly veteran population in the baseline period (2013). Next, we examined unadjusted changes in coverage type between 2013 and 2014, overall and for urban and rural veterans separately. Finally, we estimated differences in the adjusted probability of being insured (any type of coverage) between 2013 and 2014. We stratified the sample by urban and rural status and computed these differences separately for each strata. Adjusted probabilities were computed by fitting logistic regressions that included year and controls for age, gender, race, marital status, poverty status, educational attainment, and employment. Two separate versions of the model were run: one that interacted year with Medicaid expansion status and another that interacted year with income category. We used t-tests to determine statistical significance at the p < 0.05 level. Coefficient estimates were converted to a probability scale for ease of interpretation.18 More details of our statistic model can be found in the Supplementary materials. Limitations Potential limitations to our study included the constraints imposed by the ACS on how we defined urban/rural status. Because the lowest level of geography was PUMAs and PUMA boundaries might encompass both urban and rural communities, our measure might have led to misclassification (we explore this issue more thoroughly in the Supplementary material). In addition, there was a change in the way veterans were identified in the ACS starting in 2013. As a result, we were only able to examine changes in coverage between 2013 and 2014. The lack of a substantial pre-period prevented us from determining if the changes we observed were simply a continuation of long-standing trends. However, data from other sources suggest stable rates going back at least to 2011.19 Furthermore, our results are in line with a growing body of literature which finds that the ACA is associated with increased coverage and access to care, independent of other factors.20 RESULTS Table I presents baseline characteristics of non-elderly veterans by urban and rural status. Table I. Characteristics of Community-Dwelling Non-Elderly Urban and Rural Veterans, 2013   Urban  Rural  Total  %  SE  %  SE  %  SE  2013 sample size  79,816  22,739  102,555  2013 population count  8,301,388  1,800,501  10,101,889  Region   Northeast  12.85  0.13  12.89  0.27  12.85  0.12   Midwest  19.31  0.18  29.79*  0.35  21.17  0.16   South  42.48  0.18  49.16*  0.38  43.67  0.17   West  25.37  0.16  8.17*  0.25  22.30  0.14  Age   18–34  16.45  0.15  12.10*  0.29  15.67  0.13   35–44  18.11  0.18  15.86*  0.30  17.71  0.15   45–64  65.45  0.20  72.04*  0.42  66.63  0.17  Gender   Male  87.02  0.13  89.08*  0.24  87.39  0.13   Female  12.98  0.13  10.92*  0.24  12.61  0.13  Race   White, Non-Hispanic  69.29  0.24  85.49*  0.30  72.18  0.21   Black, Non-Hispanic  16.51  0.15  8.47*  0.21  15.08  0.13   Other, Non-Hispanic  5.01  0.08  3.32*  0.15  4.71  0.08   Hispanic  9.18  0.16  2.73*  0.15  8.03  0.13  Marriage   Not married  39.27  0.23  35.86*  0.45  38.66  0.19   Married  60.73  0.23  64.14*  0.45  61.34  0.19  Family income   0–138% FPG  17.02  0.16  20.00*  0.36  17.55  0.14   139–249% FPG  15.58  0.14  18.98*  0.36  16.19  0.13   250–399% FPG  21.42  0.18  24.71*  0.32  22.01  0.16   ≥400% FPG  45.98  0.23  36.32*  0.40  44.26  0.20  Educational attainment   High School, GED, or less  28.88  0.20  39.59*  0.41  30.69  0.07   Some college  44.92  0.20  42.79*  0.45  44.53  0.18   Bachelor’s or more  26.32  0.19  17.62*  0.30  24.77  0.19  Employment   Unemployed  5.67  0.11  4.95*  0.20  5.54  0.10   Part-timer worker  4.56  0.08  4.41  0.18  4.53  0.07   Full-time worker  66.75  0.20  61.04*  0.39  65.73  0.17   Not in labor force  23.03  0.18  29.61*  0.39  24.20  0.16  Functional limitations   None  82.84  0.18  77.33*  0.38  81.86  0.16   Any  17.16  0.18  22.67*  0.38  18.14  0.16  Service connected   No  78.69  0.16  78.53  0.32  78.66  0.14   Yes  21.31  0.16  21.47  0.32  21.34  0.14  Most recent period of service   Vietnam era or earlier  16.56  0.14  12.58*  0.29  15.85  0.12   1975–1990  19.91  0.15  17.04*  0.35  19.40  0.13   Gulf War Era  37.04  0.21  37.11  0.40  37.05  0.18   Since 911  26.49  0.19  33.28*  0.35  27.70  0.17    Urban  Rural  Total  %  SE  %  SE  %  SE  2013 sample size  79,816  22,739  102,555  2013 population count  8,301,388  1,800,501  10,101,889  Region   Northeast  12.85  0.13  12.89  0.27  12.85  0.12   Midwest  19.31  0.18  29.79*  0.35  21.17  0.16   South  42.48  0.18  49.16*  0.38  43.67  0.17   West  25.37  0.16  8.17*  0.25  22.30  0.14  Age   18–34  16.45  0.15  12.10*  0.29  15.67  0.13   35–44  18.11  0.18  15.86*  0.30  17.71  0.15   45–64  65.45  0.20  72.04*  0.42  66.63  0.17  Gender   Male  87.02  0.13  89.08*  0.24  87.39  0.13   Female  12.98  0.13  10.92*  0.24  12.61  0.13  Race   White, Non-Hispanic  69.29  0.24  85.49*  0.30  72.18  0.21   Black, Non-Hispanic  16.51  0.15  8.47*  0.21  15.08  0.13   Other, Non-Hispanic  5.01  0.08  3.32*  0.15  4.71  0.08   Hispanic  9.18  0.16  2.73*  0.15  8.03  0.13  Marriage   Not married  39.27  0.23  35.86*  0.45  38.66  0.19   Married  60.73  0.23  64.14*  0.45  61.34  0.19  Family income   0–138% FPG  17.02  0.16  20.00*  0.36  17.55  0.14   139–249% FPG  15.58  0.14  18.98*  0.36  16.19  0.13   250–399% FPG  21.42  0.18  24.71*  0.32  22.01  0.16   ≥400% FPG  45.98  0.23  36.32*  0.40  44.26  0.20  Educational attainment   High School, GED, or less  28.88  0.20  39.59*  0.41  30.69  0.07   Some college  44.92  0.20  42.79*  0.45  44.53  0.18   Bachelor’s or more  26.32  0.19  17.62*  0.30  24.77  0.19  Employment   Unemployed  5.67  0.11  4.95*  0.20  5.54  0.10   Part-timer worker  4.56  0.08  4.41  0.18  4.53  0.07   Full-time worker  66.75  0.20  61.04*  0.39  65.73  0.17   Not in labor force  23.03  0.18  29.61*  0.39  24.20  0.16  Functional limitations   None  82.84  0.18  77.33*  0.38  81.86  0.16   Any  17.16  0.18  22.67*  0.38  18.14  0.16  Service connected   No  78.69  0.16  78.53  0.32  78.66  0.14   Yes  21.31  0.16  21.47  0.32  21.34  0.14  Most recent period of service   Vietnam era or earlier  16.56  0.14  12.58*  0.29  15.85  0.12   1975–1990  19.91  0.15  17.04*  0.35  19.40  0.13   Gulf War Era  37.04  0.21  37.11  0.40  37.05  0.18   Since 911  26.49  0.19  33.28*  0.35  27.70  0.17  Source: 2013 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census block, as defined by the Census Bureau. SE are standard errors estimated using successive difference replication. *Indicates a statistically significant difference between urban and rural at the p< 0.05 level. Table I. Characteristics of Community-Dwelling Non-Elderly Urban and Rural Veterans, 2013   Urban  Rural  Total  %  SE  %  SE  %  SE  2013 sample size  79,816  22,739  102,555  2013 population count  8,301,388  1,800,501  10,101,889  Region   Northeast  12.85  0.13  12.89  0.27  12.85  0.12   Midwest  19.31  0.18  29.79*  0.35  21.17  0.16   South  42.48  0.18  49.16*  0.38  43.67  0.17   West  25.37  0.16  8.17*  0.25  22.30  0.14  Age   18–34  16.45  0.15  12.10*  0.29  15.67  0.13   35–44  18.11  0.18  15.86*  0.30  17.71  0.15   45–64  65.45  0.20  72.04*  0.42  66.63  0.17  Gender   Male  87.02  0.13  89.08*  0.24  87.39  0.13   Female  12.98  0.13  10.92*  0.24  12.61  0.13  Race   White, Non-Hispanic  69.29  0.24  85.49*  0.30  72.18  0.21   Black, Non-Hispanic  16.51  0.15  8.47*  0.21  15.08  0.13   Other, Non-Hispanic  5.01  0.08  3.32*  0.15  4.71  0.08   Hispanic  9.18  0.16  2.73*  0.15  8.03  0.13  Marriage   Not married  39.27  0.23  35.86*  0.45  38.66  0.19   Married  60.73  0.23  64.14*  0.45  61.34  0.19  Family income   0–138% FPG  17.02  0.16  20.00*  0.36  17.55  0.14   139–249% FPG  15.58  0.14  18.98*  0.36  16.19  0.13   250–399% FPG  21.42  0.18  24.71*  0.32  22.01  0.16   ≥400% FPG  45.98  0.23  36.32*  0.40  44.26  0.20  Educational attainment   High School, GED, or less  28.88  0.20  39.59*  0.41  30.69  0.07   Some college  44.92  0.20  42.79*  0.45  44.53  0.18   Bachelor’s or more  26.32  0.19  17.62*  0.30  24.77  0.19  Employment   Unemployed  5.67  0.11  4.95*  0.20  5.54  0.10   Part-timer worker  4.56  0.08  4.41  0.18  4.53  0.07   Full-time worker  66.75  0.20  61.04*  0.39  65.73  0.17   Not in labor force  23.03  0.18  29.61*  0.39  24.20  0.16  Functional limitations   None  82.84  0.18  77.33*  0.38  81.86  0.16   Any  17.16  0.18  22.67*  0.38  18.14  0.16  Service connected   No  78.69  0.16  78.53  0.32  78.66  0.14   Yes  21.31  0.16  21.47  0.32  21.34  0.14  Most recent period of service   Vietnam era or earlier  16.56  0.14  12.58*  0.29  15.85  0.12   1975–1990  19.91  0.15  17.04*  0.35  19.40  0.13   Gulf War Era  37.04  0.21  37.11  0.40  37.05  0.18   Since 911  26.49  0.19  33.28*  0.35  27.70  0.17    Urban  Rural  Total  %  SE  %  SE  %  SE  2013 sample size  79,816  22,739  102,555  2013 population count  8,301,388  1,800,501  10,101,889  Region   Northeast  12.85  0.13  12.89  0.27  12.85  0.12   Midwest  19.31  0.18  29.79*  0.35  21.17  0.16   South  42.48  0.18  49.16*  0.38  43.67  0.17   West  25.37  0.16  8.17*  0.25  22.30  0.14  Age   18–34  16.45  0.15  12.10*  0.29  15.67  0.13   35–44  18.11  0.18  15.86*  0.30  17.71  0.15   45–64  65.45  0.20  72.04*  0.42  66.63  0.17  Gender   Male  87.02  0.13  89.08*  0.24  87.39  0.13   Female  12.98  0.13  10.92*  0.24  12.61  0.13  Race   White, Non-Hispanic  69.29  0.24  85.49*  0.30  72.18  0.21   Black, Non-Hispanic  16.51  0.15  8.47*  0.21  15.08  0.13   Other, Non-Hispanic  5.01  0.08  3.32*  0.15  4.71  0.08   Hispanic  9.18  0.16  2.73*  0.15  8.03  0.13  Marriage   Not married  39.27  0.23  35.86*  0.45  38.66  0.19   Married  60.73  0.23  64.14*  0.45  61.34  0.19  Family income   0–138% FPG  17.02  0.16  20.00*  0.36  17.55  0.14   139–249% FPG  15.58  0.14  18.98*  0.36  16.19  0.13   250–399% FPG  21.42  0.18  24.71*  0.32  22.01  0.16   ≥400% FPG  45.98  0.23  36.32*  0.40  44.26  0.20  Educational attainment   High School, GED, or less  28.88  0.20  39.59*  0.41  30.69  0.07   Some college  44.92  0.20  42.79*  0.45  44.53  0.18   Bachelor’s or more  26.32  0.19  17.62*  0.30  24.77  0.19  Employment   Unemployed  5.67  0.11  4.95*  0.20  5.54  0.10   Part-timer worker  4.56  0.08  4.41  0.18  4.53  0.07   Full-time worker  66.75  0.20  61.04*  0.39  65.73  0.17   Not in labor force  23.03  0.18  29.61*  0.39  24.20  0.16  Functional limitations   None  82.84  0.18  77.33*  0.38  81.86  0.16   Any  17.16  0.18  22.67*  0.38  18.14  0.16  Service connected   No  78.69  0.16  78.53  0.32  78.66  0.14   Yes  21.31  0.16  21.47  0.32  21.34  0.14  Most recent period of service   Vietnam era or earlier  16.56  0.14  12.58*  0.29  15.85  0.12   1975–1990  19.91  0.15  17.04*  0.35  19.40  0.13   Gulf War Era  37.04  0.21  37.11  0.40  37.05  0.18   Since 911  26.49  0.19  33.28*  0.35  27.70  0.17  Source: 2013 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census block, as defined by the Census Bureau. SE are standard errors estimated using successive difference replication. *Indicates a statistically significant difference between urban and rural at the p< 0.05 level. Overall, we estimated 10.1 million U.S. non-elderly veterans in 2013; 82% lived in predominantly urban PUMAs (8.3 million), and the remaining 18% (1.8 million) lived in predominately rural PUMAs. Most veterans lived in the South (43.6%), and rural veterans were more likely to be Southerners than their urban counterparts (P < 0.05). On every marker of economic well-being, rural veterans fared worse than urban veterans. They had a statistically significant higher chance of having incomes being below 138% of FPG (20.0% versus 17.0%), of being out of the labor force (29.1% versus 23.0%), and of having no more than a high school education (39.6% versus 28.8%). Rural veterans were also more likely to experience at least one functional limitation, but they had the same chance of having a service connected disability compared to their urban counterparts. Table II presents unadjusted estimates of health insurance type by year, overall and separately for urban and rural veterans. Non-elderly veterans, regardless of urban/rural status, experienced a significant 2.2 percentage point reduction in uninsurance during the first year of the ACA (from 2013 to 2014). This translated into a reduction in the number of uninsured Veterans by approximately 230,000. Overall, coverage gains were driven by increases in employer-sponsored coverage (0.8 percentage points) and non-VA public coverage (primarily Medicaid) (0.8 percentage points). There was also a significant increase in the number of veterans accessing VHA care, but there was no change in the percent of veterans who only had VHA care and no other coverage type. Table II. Changes in Coverage Type for Community-Dwelling Non-Elderly Urban and Rural Vets, 2013–2014   2013  2014  Difference  %  SE  %  SE    Total sample size  102,555    96,843      Total population size  10,101,889    9,715,012      Total   Employer/tri-care coverage  70.69  0.21  71.47  0.17  0.79*   Medicaid/Medicare/other public  11.73  0.11  12.57  0.14  0.83*   Directly purchased coverage  7.70  0.11  7.97  0.11  0.28   VA coverage (Any)  27.22  0.20  28.17  0.18  0.95*   VA coverage (Alone)  8.22  0.12  8.22  0.11  0.00   Uninsured  9.31  0.12  7.11  0.11  −2.2*  Urban   Employer/tri-care coverage  71.69  0.22  72.62  0.19  0.92*   Medicaid/Medicare/other public  11.01  0.12  11.76  0.15  0.75*   Directly purchased coverage  7.68  0.12  7.85  0.12  0.17   VA coverage (any)  26.68  0.22  27.68  0.22  0.99*   VA coverage (alone)  8.04  0.14  8.11  0.12  0.07   Uninsured  9.01  0.13  6.76  0.12  −2.25*  Rural   Employer/tri-care coverage  66.06  0.40  66.18  0.44  0.12   Medicaid/Medicare/other public  15.08  0.33  16.31  0.33  1.23*   Directly purchased coverage  7.77  0.23  8.53  0.24  0.76*   VA coverage (any)  29.72  0.36  30.48  0.39  0.76   VA coverage (alone)  9.04  0.23  8.72  0.25  −0.33   Uninsured  10.73  0.29  8.74  0.25  −1.99*    2013  2014  Difference  %  SE  %  SE    Total sample size  102,555    96,843      Total population size  10,101,889    9,715,012      Total   Employer/tri-care coverage  70.69  0.21  71.47  0.17  0.79*   Medicaid/Medicare/other public  11.73  0.11  12.57  0.14  0.83*   Directly purchased coverage  7.70  0.11  7.97  0.11  0.28   VA coverage (Any)  27.22  0.20  28.17  0.18  0.95*   VA coverage (Alone)  8.22  0.12  8.22  0.11  0.00   Uninsured  9.31  0.12  7.11  0.11  −2.2*  Urban   Employer/tri-care coverage  71.69  0.22  72.62  0.19  0.92*   Medicaid/Medicare/other public  11.01  0.12  11.76  0.15  0.75*   Directly purchased coverage  7.68  0.12  7.85  0.12  0.17   VA coverage (any)  26.68  0.22  27.68  0.22  0.99*   VA coverage (alone)  8.04  0.14  8.11  0.12  0.07   Uninsured  9.01  0.13  6.76  0.12  −2.25*  Rural   Employer/tri-care coverage  66.06  0.40  66.18  0.44  0.12   Medicaid/Medicare/other public  15.08  0.33  16.31  0.33  1.23*   Directly purchased coverage  7.77  0.23  8.53  0.24  0.76*   VA coverage (any)  29.72  0.36  30.48  0.39  0.76   VA coverage (alone)  9.04  0.23  8.72  0.25  −0.33   Uninsured  10.73  0.29  8.74  0.25  −1.99*  Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census block, as defined by the Census Bureau. Coverage types are not mutually exclusive, see text for details. SE are standard errors estimated using successive difference replication. *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. Note that none of the changes over time were statistically different in rural versus urban areas. Table II. Changes in Coverage Type for Community-Dwelling Non-Elderly Urban and Rural Vets, 2013–2014   2013  2014  Difference  %  SE  %  SE    Total sample size  102,555    96,843      Total population size  10,101,889    9,715,012      Total   Employer/tri-care coverage  70.69  0.21  71.47  0.17  0.79*   Medicaid/Medicare/other public  11.73  0.11  12.57  0.14  0.83*   Directly purchased coverage  7.70  0.11  7.97  0.11  0.28   VA coverage (Any)  27.22  0.20  28.17  0.18  0.95*   VA coverage (Alone)  8.22  0.12  8.22  0.11  0.00   Uninsured  9.31  0.12  7.11  0.11  −2.2*  Urban   Employer/tri-care coverage  71.69  0.22  72.62  0.19  0.92*   Medicaid/Medicare/other public  11.01  0.12  11.76  0.15  0.75*   Directly purchased coverage  7.68  0.12  7.85  0.12  0.17   VA coverage (any)  26.68  0.22  27.68  0.22  0.99*   VA coverage (alone)  8.04  0.14  8.11  0.12  0.07   Uninsured  9.01  0.13  6.76  0.12  −2.25*  Rural   Employer/tri-care coverage  66.06  0.40  66.18  0.44  0.12   Medicaid/Medicare/other public  15.08  0.33  16.31  0.33  1.23*   Directly purchased coverage  7.77  0.23  8.53  0.24  0.76*   VA coverage (any)  29.72  0.36  30.48  0.39  0.76   VA coverage (alone)  9.04  0.23  8.72  0.25  −0.33   Uninsured  10.73  0.29  8.74  0.25  −1.99*    2013  2014  Difference  %  SE  %  SE    Total sample size  102,555    96,843      Total population size  10,101,889    9,715,012      Total   Employer/tri-care coverage  70.69  0.21  71.47  0.17  0.79*   Medicaid/Medicare/other public  11.73  0.11  12.57  0.14  0.83*   Directly purchased coverage  7.70  0.11  7.97  0.11  0.28   VA coverage (Any)  27.22  0.20  28.17  0.18  0.95*   VA coverage (Alone)  8.22  0.12  8.22  0.11  0.00   Uninsured  9.31  0.12  7.11  0.11  −2.2*  Urban   Employer/tri-care coverage  71.69  0.22  72.62  0.19  0.92*   Medicaid/Medicare/other public  11.01  0.12  11.76  0.15  0.75*   Directly purchased coverage  7.68  0.12  7.85  0.12  0.17   VA coverage (any)  26.68  0.22  27.68  0.22  0.99*   VA coverage (alone)  8.04  0.14  8.11  0.12  0.07   Uninsured  9.01  0.13  6.76  0.12  −2.25*  Rural   Employer/tri-care coverage  66.06  0.40  66.18  0.44  0.12   Medicaid/Medicare/other public  15.08  0.33  16.31  0.33  1.23*   Directly purchased coverage  7.77  0.23  8.53  0.24  0.76*   VA coverage (any)  29.72  0.36  30.48  0.39  0.76   VA coverage (alone)  9.04  0.23  8.72  0.25  −0.33   Uninsured  10.73  0.29  8.74  0.25  −1.99*  Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census block, as defined by the Census Bureau. Coverage types are not mutually exclusive, see text for details. SE are standard errors estimated using successive difference replication. *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. Note that none of the changes over time were statistically different in rural versus urban areas. The decline in the uninsured was similar for both urban and rural populations. However, the specific types of coverage driving the reduction in uninsured differed between urban and rural status. Veterans in urban areas experienced a significant increase in employer-sponsored insurance of 1 percentage point and significant increase in non-VA public coverage of 0.75 percentage points. Rural veterans did not experience a significant change in employer-sponsored insurance, but they did experience a significant gain in non-VA public coverage (1.23 percentage points). Rural veterans also experienced a statistically significant increase in direct purchase coverage while the change in direct purchase coverage for urban veterans was small and not significant. In Figures 1 and 2, we present adjusted differences in the percent of veterans that were insured between 2013 and 2014. Complete regression results are provided in the Supplementary material. Figure 1 describes changes in the percent insured by urban/rural status, for veterans living in Medicaid expansion states versus those living in non-expansion states. Overall, veterans in Medicaid expansion states experienced a significantly larger increase in insurance compared with veterans living in non-expansion states. Rural veterans in Medicaid expansion states had significantly larger increases compared to rural veterans in non-expansion states. For rural veterans in Medicaid expansion states, the increase in insurance was 3.5 percentage points, compared with 1.2 percentage points in non-expansion states. While all urban veterans experienced a significant increase in insurance, the relative increase in Medicaid expansion versus non- expansion states was considerably smaller than it was for rural veterans, and it was not significant. Figure 1. View largeDownload slide Differences in the Adjusted Percent Insured Between 2014 and 2013, by Urban/Rural and Expansion Status. Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census Block, as defined by the Census Bureau. Results represent average marginal effects from 3 separate logistic regressions (see text for details). *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. †indicates that 2014–2013 difference is significantly different in Medicaid expansion versus non-expansion states (P < 0.05). Please see the Supplementary material for complete regression results. Figure 1. View largeDownload slide Differences in the Adjusted Percent Insured Between 2014 and 2013, by Urban/Rural and Expansion Status. Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census Block, as defined by the Census Bureau. Results represent average marginal effects from 3 separate logistic regressions (see text for details). *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. †indicates that 2014–2013 difference is significantly different in Medicaid expansion versus non-expansion states (P < 0.05). Please see the Supplementary material for complete regression results. FIGURE 2. View largeDownload slide Differences in the Adjusted Percent Insured Between 2014 and 2013, by Urban/Rural and Income. Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census Block, as defined by the Census Bureau. Results represent average marginal effects from three separate logistic regressions (see text for details). *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. †Indicates that 2014–2013 difference is significantly different in the given income category compared to the lowest-income group (P < 0.05). Please see the Supplementary material for complete regression results. FIGURE 2. View largeDownload slide Differences in the Adjusted Percent Insured Between 2014 and 2013, by Urban/Rural and Income. Source: 2013–2014 American Community Survey. Rural status is defined as living in a Public Use Microdata Area in which more than 50% of the population lives in a rural Census Block, as defined by the Census Bureau. Results represent average marginal effects from three separate logistic regressions (see text for details). *Indicates a statistically significant difference between 2013 and 2014 at the p < 0.05 level. †Indicates that 2014–2013 difference is significantly different in the given income category compared to the lowest-income group (P < 0.05). Please see the Supplementary material for complete regression results. In Figure 2, we report adjusted differences by urban/rural and family income categories. Overall, those in the lowest-income group experienced the largest percentage point increase in insurance. For rural veterans, the lowest-income group had the largest increases followed by those in the 251–399% FPG group. The change for the 251–399% FPG group was not statistically different from the change in the 0–138% FPG group. We conducted additional subgroup analyses and explored how sensitive our results were to various modeling approaches (see Supplementary material). For example, because the ACA targeted those without access to employer-sponsored insurance, we examined regression adjusted changes in insurance for low-income urban and rural veterans by employment status. Non-working urban veterans experienced a 6.8 percentage point increase in insurance and non-working rural veterans experienced an 8.4 point increase. Differences for working veterans were smaller in both urban and rural communities. We also explored how sensitive our main results were to alternative modeling assumptions. We found that removing the states that were early adopters of the Medicaid expansion (expanding prior to January 1, 2014) had little impact on our Medicaid expansion results. We also obtained similar results when our Medicaid expansion regressions included state fixed effects. Finally, we explored how our results changed as we lowered the threshold at which we considered a veteran to reside in a predominantly rural community from 50% to 25%. Lowering the threshold had no effect on estimates of declines in uninsurance by urban and rural status, but as we reduced the threshold we found that changes to specific coverage types tended to appear similar across rural and urban areas. This did not surprise us, as lowering the threshold had the effect of increasing the “urban-ness” of those PUMAs we initially classified as predominantly rural. CONCLUSION There was a significant increase in insurance for U.S. veterans during the first year of the ACA. Our results suggest a substantial 24% relative decline in uninsurance from 2013 to 2014 – from a 9.3% uninsured rate to 7.1%. The changes in insurance status we observed were similar for both urban and rural veterans, but coverage gains by rural/urban status were driven by different coverage pathways. In urban areas, the increase in coverage was driven in large part by gains in employer-sponsored insurance that was likely the result of an improving economy and not the direct effect of the ACA. In rural areas, the increase in coverage was largely due to gains in Medicaid and individual market coverage. Residence in a Medicaid expansion state was particularly influential for rural veterans – the decline in the uninsured rate was three times larger in Medicaid expansion states versus non-expansion states. Our analysis suggests that Medicaid expansion was associated with a significant increase in health insurance coverage rates for rural veterans; we did not find a similar association for urban veterans. Rural veterans had neither gains nor losses in employer-sponsored insurance. Relatedly, we found that rural veterans were more likely to be out of the labor force compared to urban veterans and thus Medicaid may have played an important role in buffering the negative impact of a weak labor market. Two-thirds of non-elderly U.S. veterans were either Gulf War veterans or in service since 9/11, between the ages of 45 and 64, and working full time. Not surprisingly, most veterans (71.5% in 2014) obtained their health insurance through their or a spouse’s place of employment, higher than the rate for the non-elderly with employer-sponsored insurance for the general U.S. population (56%).21 Yet for low-income veterans including those without jobs or employer-sponsored insurance, the ACA has provided new pathways for obtaining health insurance coverage which have supplements services and financing provided through the VHA. We found that the Medicaid expansion was a particular important source of coverage for rural veterans. A concern at this point is for the many (43.7%) veterans living in the south where 10 out of 19 states (as of 2016) that have not expanded Medicaid22 are located.23 U.S. veterans rely on both public and private health insurance coverage in addition to VHA coverage. As policymakers and VA administrators continue to seek and experiment with new policy alternatives for increasing access to health services for veterans, the new coverage options created by the ACA will be an important tool. Perhaps more importantly, states that have not yet expanded Medicaid should consider the impact that expansion would have on increased coverage and access to care for low-income veterans and particularly those residing in rural areas. Supplementary Material Supplementary material is available at Military Medicine online. Acknowledgement This study was funded in part under a grant from the Robert Wood Johnson Foundation to the State Health Access Data Assistance Center at the University of Minnesota School of Public Health. Funding This study was funded in part under a grant from the Robert Wood Johnson Foundation to the State Health Access Data Assistance Center at the University of Minnesota School of Public Health. References 1 VA health care 2016: Developing solutions for the next generation of veteran care. International Quality & Productivity Center. Available at http://www.veteransaffairshealthcare.com/; accessed March 5, 2016. 2 Choski DA, Sommers B: Universal health coverage for US veterans: a goal within reach. Lancet  2014; 385( 9984): 2320– 1. Google Scholar PubMed  3 Haley J, Kenney GM; The Urban Institute. Uninsured veterans and family members: state and national estimates of expanded Medicaid eligibility under the ACA. Available at http://research.urban.org/UploadedPDF/412775-Uninsured-Veterans-and-Family-Members.pdf. 2013; accessed March 5, 2016. 4 Federal Register (US). Annual update of the HHS Federal Poverty Guidelines. 2014;70 FR(3593):2593–4. Available at https://federalregister.gov/a/2014-01303. 2014; accessed August 24, 2016. 5 Henry J Kaiser Family Foundation. Status of state action on the Medicaid expansion decision. Available at http://kff.org/health-reform/state-indicator/state-activity-around-expanding-medicaid-under-the-affordable-care-act/. 2016; accessed March 5, 2016. 6 Sommers BD, Gunja M, Finegold K, et al.  : Changes in self-reported Insurance coverage, access to care, and health under the affordable care act. JAMA.  2015; 314( 4): 366– 74. Google Scholar CrossRef Search ADS PubMed  7 Sommers BD, Maylone B, Nguyen KH, et al.  : The impact of state policies on ACA applications and enrollment among low-income adults in Arkansas, Kentucky, and Texas. Health Affairs  2015; 34( 6): 1010– 8. Google Scholar CrossRef Search ADS PubMed  8 Chen J, Vargas-Bustamante A, Mortensen K, et al.  : Racial and ethnic disparities in health care access and utilization under the Affordable Care Act. Med Care  2016; 54( 2): 140– 6. Google Scholar CrossRef Search ADS PubMed  9 Shartzer A, Long SK, Karpman M, et al.  : Quicktake: Insurance coverage gains cross economic, social, and geographic boundaries. Health Reform Monitoring Survey. Available at http://hrms.urban.org/quicktakes/Insurance-Coverage-Gains-Cross-Economic-Social-and-Geographic-Boundaries.html. 2015; accessed March 5, 2016. 10 Drake C, Abraham JM, McCullough JS: Rural enrollment in the federally facilitated marketplace. J Rural Health  2016; 32( 3): 332– 9. Google Scholar CrossRef Search ADS PubMed  11 Gooptu A, Moriya AS, Simon KI, et al.  : Medicaid expansion did not result in significant employment changes or job reductions in 2014. Health Aff  2016; 35( 1): 111– 8. Google Scholar CrossRef Search ADS   12 Haley J, Kenney GM; The Urban Institute. Uninsured veterans and family members: who are they and where do they live? Available at http://www.urban.org/research/publication/uninsured-veterans-and-family-members-who-are-they-and-where-do-they-live. 2012; accessed March 5, 2016. 13 Department of Veterans Affairs. About rural veterans. Available at http://www.ruralhealth.va.gov/about/rural-veterans.asp; accessed March 5, 2016. 14 Buzza C, Ono SS, Turvey C, et al.  : Distance is relative: unpacking a principal barrier in rural healthcare. J Gen Intern Med  2011; 26( Suppl 2): 648– 54. doi:10.1007/s11606-011-1762-1; accessed March 5, 2016. Google Scholar CrossRef Search ADS PubMed  15 The United States Census Bureau. American Community Survey response rates. Available at http://www.census.gov/acs/www/methodology/sample-size-and-data-quality/response-rates/; accessed March 5, 2016. 16 Missouri Census Data Center. MABLE/Geocorr12, Version 12: Geographic correspondence engine. Available at http://mcdc.missouri.edu/websas/geocorr12.html; accessed February 2016. 17 Henry J Kaiser Family Foundation. Medicaid eligibility for adults as of January 1, 2014. Available at http://kff.org/medicaid/fact-sheet/medicaid-eligibility-for-adults-as-of-january-1-2014/. 2013 accessed March 5, 2016. 18 Karaca-Mandic P, Norton E, Dowd B: Interaction terms in nonlinear models. Health Serv Res  2012; 47( 1.1): 255– 74. Google Scholar CrossRef Search ADS PubMed  19 Haley J, Kenney G; The Urban Institute. The Veterans Day, early evidence that the ACA is serving those who have served. Available at http://www.urban.org/urban-wire/veterans-day-early-evidence-aca-serving-those-who-have-served. 2015; accessed March 5, 2016. 20 French MT, Homer J, Gumus G, et al.  : Key provisions of the Patient Protection and Affordable Care Act (ACA): A systemic review and presentation of early research findings. Health Serv Res  2016; 51( 5): 1735– 71. doi:10.1111/1475-6773.12511; accessed March 5, 2016. Google Scholar CrossRef Search ADS PubMed  21 Henry J Kaiser Family Foundation. Health insurance coverage of nonelderly 0–64. Available at http://kff.org/other/state-indicator/nonelderly-0-64/?currentTimeframe=0&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D; accessed August 30, 2016. 22 Families USA. A 50-state look at Medicaid expansion. Available at http://familiesusa.org/product/50-state-look-medicaid-expansion. 2016; accessed August 30, 2016. 23 U.S. Census Bureau. Census Bureau regions and divisions with state FIPS codes. Available at https://www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf; accessed August 30, 2016. Author notes The views expressed are those of the authors and do not necessarily represent the views of the Robert Wood Johnson Foundation. © Association of Military Surgeons of the United States 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Military MedicineOxford University Press

Published: Apr 25, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off