Affordable Care Act Impact in Kentucky: Increasing Access, Reducing Disparities

Abstract

Objectives. To examine health insurance disparities since Kentucky’s implementation of the Affordable Care Act (ACA).

Methods. Using the American Community Survey, we estimated coverage rates by race/ethnicity before and after implementation of the ACA (2013 and 2015), and we estimated whether groups were over- or underrepresented among the uninsured, compared with their share of the state population.

Results. Kentucky’s uninsurance rate declined from 14.4% in 2013 to 6.1% in 2015 (P < .001). Uninsurance rates also declined for most racial/ethnic groups, including Blacks (16.7% to 5.5%; P < .001) and Whites (13.3% to 5.3%; P < .001). In 2015, Blacks were no longer overrepresented among Kentucky’s uninsured, with a significant decline in the ratio of Blacks among the state uninsured population compared with their share of the state population (1.16–0.91; P = .045).

Conclusions. In Kentucky, which mounted a robust implementation of the ACA—including Medicaid expansion, a state-based marketplace, and an extensive outreach and enrollment campaign—the state experienced not only a decline in the overall uninsurance rate but also an elimination in coverage disparities among Blacks, who historically were overrepresented among the uninsured.

Kentucky has received national attention for its implementation of the Affordable Care Act (ACA). Despite disagreements within Kentucky over the extent to which the state should embrace the ACA, the state’s Democratic governor, Steve Beshear, used his executive authority to establish a State-Based Marketplace (SBM), formerly called “kynect”; expand its state Medicaid program to the ACA’s newly eligible adult population; and undertake an extensive and innovative outreach and enrollment effort, including a mobile phone app and a retail location in a busy shopping mall.^1,2 In the first year of ACA implementation, Kentucky experienced the largest decline in uninsurance rates in the United States—falling from 14.4% in 2013 to 8.7% in 2014.³ The most recent estimates show Kentucky with an overall uninsurance rate of 5.1% in 2016. This article represents an analysis grown from a more-comprehensive study of the impact of the implementation of the ACA in Kentucky.² Here we document increases in health insurance coverage across nearly all population groups and the elimination of Kentucky’s significant and persistent coverage disparities between the White and Black residents within the first 2 years of ACA implementation.

There is a growing body of research demonstrating the impact of the ACA on increasing access to health insurance coverage and on reducing historical disparities in access to coverage by race/ethnicity.^4–6 Following the first 2 years of implementation of the ACA, the uninsured rate dropped 9 percentage points for Black nonelderly adults (aged 19–64 years) and by 12 percentage points for Hispanic nonelderly adults while the uninsured rate for White nonelderly adults declined by 5 percentage points.⁷ Buchmueller et al. examined the first year of ACA implementation using data from the American Community Survey (ACS) from 2008 to 2014 and found a reduction in coverage disparities related to race/ethnicity under the ACA, with both Blacks and Hispanics experiencing greater gains in coverage than Whites experienced.⁸ Using data from the ACS and the Behavioral Risk Factor Surveillance System, Hayes et al. also found reduced racial/ethnic disparities in access to coverage with greater improvement in the Medicaid expansion states.⁹

Other studies have found that increased access to coverage also led to increases in access to needed care and decreases in cost-related barriers to care. Sommers et al. found that, in addition to higher rates of health insurance coverage, non-Whites in Medicaid-expansion states Kentucky and Arkansas experienced significant increases in office visits with health care providers and improvements in multiple measures of affordability of care.¹⁰ Hayes et al. found that since the implementation of the ACA, coverage disparities narrowed between White nonelderly adults and Hispanic and Black nonelderly adults, as did disparities in measures of financial barriers (forgone care because of cost) and access to care (having a usual source of care).⁷

Our contribution to this research was to document the impact of the ACA on eliminating coverage disparities between Blacks and Whites in Kentucky and to highlight the strategies employed that may have contributed to that result. We conclude with some caution about recent changes in the state and federal policy arenas that may threaten the sustainability of Kentucky’s coverage gains.

METHODS

We used a cross-sectional observational analysis, before and after ACA implementation, with data from the US Census Bureau’s ACS. The ACS provided a representative sample of noninstitutionalized Kentuckians (all ages; n = 43 475 in 2013 and 43 455 in 2015). The ACS is conducted throughout each year, and health insurance is measured at a point in time (i.e., respondents’ health insurance coverage at the time of the survey). Individuals are considered insured if they reported having at least 1 form of coverage—private coverage (e.g., employer-based plan and individual-market coverage) or public coverage (e.g., Medicaid, Children’s Health Insurance Program [CHIP], Medicare)—and uninsured if they have neither.

Our analysis was based on comparisons of weighted rates of coverage in 2013 and 2015. We tested for statistical significance of these comparisons by using t test of weighted means in which the null hypothesis was that the difference between the 2 estimates was statistically equal to zero. In addition to comparing rates of uninsurance, we compared the racial/ethnic distribution among the uninsured with that of the overall population in each year. In a situation with no racial/ethnic disparities in uninsurance, these 2 population distributions would be expected to be equal. If a racial/ethnic group was overrepresented among the uninsured, it could be concluded that this group faced an unfavorable disparity. Conversely, a favorable disparity could be concluded from finding that 1 group was underrepresented among the uninsured.

To assess changes in disparities over time, we created a disproportionality ratio for each racial/ethnic group—the share a group represents among the uninsured divided by the group’s share in the population. We used the disproportionality ratios to evaluate for statistical significance of changes in overrepresentation (> 1.0) or underrepresentation (< 1.0) of groups among the uninsured.¹¹

RESULTS

Between 2013 and 2015, Kentucky’s uninsurance rate for all ages declined a statistically significant 8.3 percentage points, from 14.4% to 6.1% (Table 1). Our analysis by racial/ethnic groups showed that the largest decline in the uninsurance rate was experienced among Blacks—from 16.7% to 5.5% (11.2 percentage points), reaching a rate in 2015 that was not significantly different from that of the state’s White population, which also declined significantly from 13.3% in 2013 to 5.3% in 2015. Uninsurance among the Hispanic/Latino population also declined significantly (from 34.5% to 24.2%), as well as for people of other or multiple races (from 19.3% to 8.2%). Only the uninsurance rate for Asians remained statistically unchanged (17.1% in 2013 vs 12.5% in 2015—not a statistically significant difference).

TABLE 1—

Kentucky Uninsurance Rates by Race/Ethnicity, All Ages: 2013–2015

	Uninsurance Rates
Race/Ethnicity	2013, %	2015, %	P
White	13.3	5.3	< .001
Black	16.7	5.5	< .001
Asian	17.1	12.5	.15
Hispanic	34.5	24.2	< .001
Other	19.3	8.2	< .001
Total	14.4	6.1	< .001

Source. American Community Survey, public use microdata sample.

Uninsurance Disparities

The fact that White and Black Kentuckians reported the same uninsurance rate in 2015 hints at the elimination of previous disparities between these 2 groups. We further explored the racial/ethnic distribution of the uninsured and found that Blacks were significantly overrepresented among the uninsured in 2013, accounting for 8.9% of the state’s uninsured while representing a smaller 7.7% of the state’s population (P = .029; Table 2). Yet, by 2015, Blacks were no longer overrepresented among the uninsured, representing 7.1% of the state’s uninsured, not significantly different from their overall proportion of the population (7.8%; P = .384). By contrast, Whites in Kentucky were underrepresented among the uninsured in both 2013—85.9% of the population and 79.2% of the uninsured (P < .001)—and 2015—85.4% of the population and 74.2% of the uninsured (P < .001). These results show that while the unfavorable disparities for Blacks in Kentucky were eliminated between 2013 and 2015, Whites continued to experience favorable disparities.

TABLE 2—

Kentucky Disparity and Disproportionality Analysis, Affordable Care Act Impact on Uninsurance Rates, All Ages: 2013–2015

	2013 Uninsurance Disparities			2015 Uninsurance Disparities			Uninsurance Disproportionality Ratio
Race/Ethnicity	Share of Population (A), %	Share of Uninsured (B), %	P	Share of Population (C), %	Share of Uninsured (D), %	P	2013 (B/A)	2015 (D/C)	P
White	85.9	79.2	< .001	85.4	74.2	< .001	0.92	0.87	.002
Black	7.7	8.9	.029	7.8	7.1	.38	1.16	0.91	.045
Asian	1.2	1.4	.26	1.4	2.8	.007	1.18	2.06	.035
Hispanic	3.2	7.7	< .001	3.3	13.0	< .001	2.40	3.98	< .001
Other	2.1	2.8	.039	2.2	2.9	.11	1.34	1.34	.99
Total	100.0	100.0	. . .	100.0	100.0	. . .	1.00	1.00	. . .

Source. American Community Survey, public use microdata sample.

Although uninsurance rates declined significantly in Kentucky’s Hispanic/Latino population, they remained overrepresented among the state’s uninsured. In 2013, they made up 3.2% of the state population and a significantly higher 7.7% of the uninsured (P < .001), but in 2015 Hispanics/Latinos accounted for 13.0% of the state’s uninsured, compared with 3.3% of the total state population (P < .001). However, disparities changed among Asians and people of multiple or other races. In 2013, Asians in Kentucky were neither over- nor underrepresented among the uninsured, making up 1.2% of the overall population and 1.4% of the uninsured (P = .263). But in 2015, Asians were overrepresented, accounting for 1.4% of the population but a significantly higher 2.8% of the uninsured (P = .007). By contrast, Kentuckians of multiple or other races were overrepresented among the uninsured in 2013, making up 2.1% of the population but a significantly higher 2.8% of the uninsured (P = .039). In 2015, they were neither over- nor underrepresented, accounting for 2.2% of the population and 2.9% of the uninsured (P = .110).

Changes in Disparities

Through a disproportionality analysis, we tested whether these changes in over- and underrepresentation were statistically significant. In only 2 years, from 2013 to 2015, the ratio of uninsured Black Kentuckians as a share of Kentucky’s total Black population (i.e., insured and uninsured) declined from 1.16 to 0.91—a statistically significant decline (P = .045). During the same period, the ratio of uninsured White Kentuckians as a share of all White Kentuckians also declined, from 0.92 to 0.87—also a statistically significant decline (P = .002). Despite the fact that nearly all racial/ethnic groups had a decline in their uninsurance rates, only these 2 groups experienced a decline in the disproportionality ratio (and disparities) between 2013 and 2015. People who reported having multiple or other races had no significant change in their overrepresentation, remaining at a ratio of 1.34 (P = .991).

By contrast, the disproportionality ratio for Asians increased significantly, from 1.18 in 2013 to 2.06 in 2015 (P = .035), meaning that the change from Kentucky’s Asian population being neither over- nor underrepresented among the uninsured in 2013 to being overrepresented among the uninsured in 2015 was significant. Kentucky’s Latino/Hispanic population experienced the largest increase in their disproportionality ratio among all groups, from 2.40 in 2013 to 3.98 in 2015 (P < .001), representing an increase in their overrepresentation among the uninsured. These results indicate that despite the decline in uninsurance rates among Hispanics/Latinos and the statistically stable uninsurance rate among Asians, their disparities increased because other groups had even larger declines in uninsurance during the period of analysis.

DISCUSSION

While Kentucky has received national attention for achieving one of the largest declines in uninsurance in the United States under the ACA, the state’s progress in reducing some coverage disparities is less well-known. Before 2014, Black residents of Kentucky were overrepresented among the state’s uninsured. But the state’s implementation of the ACA’s coverage expansion options reduced the rate of uninsurance for Black Kentuckians to the same level as that of the White population (5.5% vs 5.3%). Black Kentuckians also were no longer overrepresented among the uninsured. Yet, reductions in coverage disparities were not universal; we did not find significant improvements in uninsurance disparities among Hispanics/Latinos, Asians, and people of other or multiple races.

Role of Policy in Reducing Disparities

A key factor in the elimination of health insurance coverage disparities among Blacks in Kentucky was that Blacks were disproportionately represented in the targeted groups for ACA expansion—the low-income and uninsured.¹² Because Kentucky’s Black population was disproportionally lower-income, the state’s decision to expand its Medicaid program disproportionally affected Black Kentuckians. In 2015, approximately half of Black Kentuckians (49.7%) had incomes at or below 138% of the federal poverty guideline (FPG) compared with just a third (33.2%) of the total Kentucky population (Blewett et al., unpublished analysis of 2015 ACS public use microdata sample). In addition, before the ACA, there were very few public options for health insurance coverage for low-income adults. In 2013, Kentucky’s income eligibility thresholds for nonelderly adults were low compared with those in many other states—33% of FPG for jobless parents and 57% FPG for working parents, and childless adults were ineligible.¹³ When Kentucky expanded its Medicaid program under the ACA, these separate eligibility categories were effectively eliminated and all categories of adults with incomes less than 138% of the FPG became eligible for Medicaid. And, as described earlier, Blacks had significantly higher uninsurance rates than did Whites in Kentucky before the ACA—a pattern that was similar across the United States.⁹

We also found that Kentucky’s Hispanic/Latino population did not experience a decline in disparities, despite having disproportionally high uninsurance rates and lower incomes (48.7% had incomes of 138% FPG or lower in 2015)—similar to Blacks (Blewett et al., unpublished analysis of 2015 ACS public use microdata sample). The continued disparities among Hispanics/Latinos are likely related to how the ACA treats immigration status in determining eligibility coverage. Nationally, nearly half of uninsured Hispanics/Latinos are estimated to be unauthorized immigrants, who are prohibited from enrolling in Medicaid or purchasing coverage through health insurance marketplaces.¹⁴ So, even if every eligible person in Kentucky enrolled in coverage, uninsurance rates among Hispanics/Latinos would remain higher. Other factors related to immigration status may also reduce enrollment in coverage (i.e., take-up) among Hispanics/Latinos who are eligible. For example, citizens and other lawfully present immigrants may be hesitant to enroll in health insurance if they have relatives in the country without authorization because of concerns about drawing attention to their families by interacting with government programs.¹⁵

In addition to Kentucky’s decision to expand Medicaid, the state made other policy changes in accordance with the ACA that likely contributed to the reduced disparities found in our study. Kentucky embraced the ACA’s “no wrong door” vision for connecting individuals with the appropriate coverage options, establishing its own SBM with an integrated eligibility and enrollment system that served individuals seeking both Medicaid and Qualified Health Plans. Enrollment was streamlined with a single common application that could be accessed in person, by phone, online, and through the mail, and a single telephone help line was available to serve all applicants. Kentucky’s SBM also largely avoided the technological problems that were experienced by other states and the federal government when setting up its marketplace.¹⁶ Then-Governor Beshear also made outreach and enrollment a top priority of his administration, providing visibility and leadership to develop a culture of coverage that encouraged people throughout the state to obtain health insurance.¹⁷

Kentucky built its outreach and enrollment work from its previous experience running extensive outreach and enrollment for its Kentucky Children’s Health Insurance Program (KCHIP) dating to the late 1990s and early 2000s—which included multimedia advertising campaigns, local outreach, and targeted efforts to enroll Hispanic/Latino populations—that was successful enough for the state’s program to exceed enrollment projections.¹⁸ Similar to the findings in our study, evidence suggests that Kentucky’s experience with coverage expansions paired with outreach and enrollment efforts among children can help to reduce coverage disparities; 2014 estimates found that uninsurance rates among White and non-White children in Kentucky were similar (4.2% White and 4.0% non-White).¹⁹ Kentucky’s experience with a reduction in some disparities following implementation of the ACA appears to mirror its earlier experience with KCHIP.

Consistent with the ACA, the state also made changes to its Medicaid application and enrollment policies, such as eliminating face-to-face interviews for adults enrolling in Medicaid, which may have acted as a barrier to take-up among people who were eligible before the ACA. Finally, Kentucky used its $6 million in outreach and enrollment funds to leverage community relationships and community-based enrollment activities that targeted populations with higher uninsurance rates. The state targeted low-income areas and used enrollment brokers (called “kynectors”) with personal ties to their communities to provide needed one-on-one enrollment assistance.²⁰ Other studies, such as those referenced in our introduction, have found reductions in coverage disparities since the ACA, especially in Medicaid expansion states, but Kentucky’s concerted effort to encourage enrollment generally and among groups with historically higher uninsured rates may help to explain why the state was an extreme case, experiencing an elimination of uninsurance disparities among Blacks.

Potential Effects of Policy Changes

Because of policy changes and uncertainty at the national and state level, there are questions about whether the state’s progress in reducing overall uninsurance rates and disparities can be sustained. There have been efforts at both the federal and state level to repeal or modify parts of the ACA. In 2017, multiple attempts were made by the US Congress to repeal and replace portions of the ACA. Although the most-comprehensive repeal bills did not pass, there have been incremental administrative and legislative changes made that may reduce the number of people with health insurance. These efforts include the decision by the US Department of Health and Human Services to cut funding for enrollment and outreach from $100 million to $10 million between 2016 and 2017, and a decision to reduce the open enrollment period from 60 to 45 days. Another significant policy change was the elimination of the tax penalty associated with the individual mandate that passed Congress as part of the tax reform bill of 2017.²¹

The Commonwealth of Kentucky has also gone through significant change. In 2016, under newly elected Republican Governor Matt Bevin, Kentucky transitioned from its kynect SBM with its own technology platform to a federally supported marketplace that uses the federal healthcare.gov enrollment platform. During the first open enrollment period under the new healthcare.gov platform, the state experienced a decline in selection of Qualified Health Plans through the marketplace, from 94 000 during the 2015–2016 enrollment period to 81 000 during the 2016–2017 enrollment period—the lowest since Kentucky’s first open-enrollment period, when kynect saw 83 000 plan selections.^22,23 It is not clear whether the decline in enrollment in private health plans through Kentucky’s marketplace may increase disparities or have little effect on them.

In January 2018, Kentucky also received federal approval to modify its Medicaid expansion program via a Section 1115 waiver, called Kentucky Helping to Engage and Achieve Long Term Health, or “HEALTH.” The waiver includes provisions allowing for premiums up to 4% income for the Medicaid expansion population; disenrollment and lock-out periods for up to 6 months for failure to pay premiums, renew, or report change in circumstances; deductible and incentive accounts similar to health savings accounts; and a first-of-its-kind work requirement. The state’s waiver is planned for phased implementation beginning July 2018 and will require most nonelderly adult Medicaid beneficiaries to participate in 80 hours per month of paid or volunteer work or other qualified activities (e.g., job training or education) or they would have their coverage suspended. In its application, Kentucky projected its waiver would result in reduced enrollment in its Medicaid program over time, from a decrease of 238 000 member months in the first year of the waiver to a decrease of 1 140 000 member months by the fifth year of the waiver.²⁴ For the same reasons that Kentucky’s implementation of the ACA decreased some coverage disparities, a decline in Medicaid enrollment caused by Kentucky’s 1115 waiver is likely to result in increased disparities. For example, because Black Kentuckians are disproportionally eligible for Medicaid under ACA expansion of the program, it is likely that Blacks will disproportionally lose Medicaid coverage as enrollment declines. In addition, as Medicaid beneficiaries are disenrolled from the program for failure to meet requirements under the waiver (e.g., payment of premiums, satisfaction of work requirements), pre-ACA patterns of coverage disparities suggest that Black Kentuckians who are disenrolled may be more likely to remain uninsured than will White Kentuckians.

As discussion continues around efforts to modify the ACA or repeal and replace the law nationally and in states, policymakers should understand that some of these changes can be expected to erode the significant gains that have been made in Kentucky on increasing access and reducing coverage disparities.

Public Health Implications

The reductions in coverage disparities that we found are encouraging, but health insurance ultimately derives its public health value as a tool for removing financial barriers and enhancing access to health care. Although our study did not examine access to and affordability of health care, other research has found improvements in those areas since implementation of the ACA’s coverage expansions. In a related earlier study of the effects of implementation of the ACA on Kentucky, we found early evidence that access to care was improving and financial barriers were declining. For example, the percentages of Kentuckians reporting a usual source of care and having a visit with a health care provider increased significantly, and the percentages of Kentuckians reporting trouble paying medical bills and delaying or forgoing care because of cost declined significantly.² Those results were consistent with a similar set of studies conducted by other researchers in Kentucky and Arkansas, which also found improvements associated with ACA Medicaid expansions, including significantly increased access to health care providers and preventive care; decreased financial barriers to care, such as skipping medications because of cost and reduced out-of-pocket spending on care; and improvements in self-reported health status.^10,25,26

The reductions in disparities we found in Kentucky demonstrate that disparities are not necessarily intractable and that public policy can be effective in mitigating them.²⁷ By improving access to and affordability of health care, reducing health coverage disparities holds potential to help enhance health equity. Although there are other critical barriers to health equity beyond health insurance coverage and access to care—including social determinants of health, such as access to transportation, food security, and housing stability—reductions in coverage disparities mark measurable progress.⁹ As Kentucky implements changes to its relationship with the ACA, it will be important to monitor health insurance coverage and disparities to understand the impacts on health equity.