Abstract
The Affordable Care Act (ACA) expanded Medicaid eligibility to adults with incomes under 138% of the federal poverty level, leading to substantial reductions in uninsured rates among low-income adults. Despite large gains in coverage, studies suggest that Latinos may be less likely than other racial/ethnic groups to apply and enroll in health insurance, and they remain the group with the highest uninsured rate in the United States. We explore two potential factors related to racial/ethnic differences in ACA enrollment—awareness of the law and receipt of application assistance such as navigator services. Using a survey of nearly 3000 low-income U.S. citizens (aged 19-64) in 3 states in late 2014, we find that Latinos had significantly lower levels of awareness of the ACA relative to other groups, even after adjusting for demographic covariates. Higher education was the strongest positive predictor of ACA awareness. In contrast, Latinos were much more likely to receive assistance from navigators or social workers when applying, relative to other racial/ethnic groups. Taken together, these results highlight the importance of ACA outreach efforts to increase awareness among low-income and less educated populations, two groups that are overrepresented in the Latino population, to close existing disparities in coverage.
Introduction
Starting in 2014, the Affordable Care Act (ACA) expanded Medicaid eligibility to adults in participating states with incomes up to 138% of the federal poverty level (FPL). Level of awareness of the ACA is a key predictor of whether people apply for coverage. 1 Despite significant outreach efforts, overall knowledge about the law remains poor among low-income and uninsured adults. 2 Working with a navigator or receiving other application assistance has been shown to be a major factor in improving enrollment rates. 1 Meanwhile, several studies have reported that Latinos may be less likely to apply and enroll in health insurance under the ACA, even though they remain the highest risk group to be uninsured in the United States.3,4 Possible explanations include lower levels of health insurance literacy, language barriers, and immigration status. 5 Our objective was to explore whether ACA awareness and receipt of application assistance differed across racial and ethnic groups among low-income U.S. citizens in 3 Southern states with high pre-ACA uninsured rates.
Methods
Our data come from a telephone survey of US citizens aged 19 to 64 with household incomes less than 138% of FPL in Texas, Arkansas, and Kentucky. The survey was conducted in November and December 2014 on mobile and landline phones, and was available in English and Spanish. Survey estimates were weighted to Census demographic benchmarks for age, sex, race/ethnicity, education, marital status, population density, and cell phone status; details of the survey have been published previously. 6 The interview focused on respondents’ awareness of and experiences with the ACA.
We used a multivariate logistic regression model to identify predictors of having heard “a lot” or “some” regarding the ACA (compared with “little” or “none”), and, among applicants, whether they received assistance from “navigators or social workers,” a phrase meant to encompass a range of application assisters, including certified application counselors and navigators, who are state-trained individuals tasked with providing application assistance and educational outreach. The key variable of interest was race/ethnicity, in the following mutually exclusive categories: Latino, white non-Latino, black non-Latino, and Other (which included those reporting more than one race or providing no response). We estimated differences by race/ethnicity after controlling for state, age, sex, education, income, rural versus urban residence, and whether the survey was completed in Spanish. We reported odds ratios and predicted probabilities associated with each covariate.
Results
Table 1 presents racial/ethnic differences in ACA awareness and use of application assistance. In the unadjusted model (n = 2794), Latinos had significantly lower awareness of the ACA than whites (P = .007), and this difference remained highly significant even after adjustment. Overall, Latinos had the lowest level of awareness of the racial/ethnic groups in our sample, with only 38.5% having heard or read “some” or “a lot” about the ACA, compared with 48% for whites and 46% for blacks, and 50.1% for Other. The fully adjusted model demonstrated that education was a much stronger predictor of ACA awareness than race/ethnicity. Respondents who attended or graduated college had nearly a 20 percentage point higher ACA awareness than respondents without high school degrees.
Predictors of Awareness of the ACA and Application Assistance Among Low-Income Adults.
Note. Total sample size = 2801. Analysis of ACA awareness excluded 7 individuals who chose not to answer that item (n = 2794). Only respondents who said they had applied for Medicaid or marketplace coverage were asked whether they received application assistance (n = 1107). Unadjusted models controlled for race/ethnicity. Full models added state of residence, age, gender, Spanish language survey, rural versus urban residence, education level, and income as covariates. We calculated predicted probabilities from the logistic regression estimates using Stata’s “margins” command with default settings, which holds all covariates at their actual values. ACA = Affordable Care Act; OR = odds ratios from multivariate logistical regression; Pred. prob. = predicted probability; FPL = federal poverty level.
P < .10. **P < .05. ***P < .01.
Table 1 also presents findings on the use of application assistance among applicants for Medicaid or Marketplace coverage (n = 1107). The unadjusted model showed no significant differences in rates of application assistance by race/ethnicity, but after adjustment, receipt of assistance was 13 percentage points higher among Latinos than whites (P = .09). Kentucky and Arkansas respondents received assistance at higher rates (48.0% and 37.3%, respectively) than respondents in Texas (28.7%).
Given that language and ethnicity are highly correlated, we tested the effect of excluding Spanish language as a covariate. The results for Latino ethnicity were similar (awareness odds ratio [OR] = 0.73, P = .07; application assistance OR = 1.81, P = .05) to those in our full model.
Discussion
In our survey of low-income US citizens in 3 states, we found that Latinos have significantly lower ACA awareness even after adjusting for income and education. ACA awareness is the critical first step in improving uninsured rates among vulnerable populations, and our findings suggest that lack of information is a key barrier among Latinos. 7 Somewhat surprisingly, language did not contribute to these informational barriers in our study, as completing the survey in Spanish was not a significant predictor of awareness. Education and race/ethnicity were more important factors, suggesting the need for targeted approaches that can reach individuals who may have low levels of experience with the health care system.
Our findings also suggest that application assistance can play an important role in reducing racial/ethnic gaps in enrollment rates. These results are consistent with previous research, especially in the case of Latinos, who are the least likely to apply for coverage, but simultaneously use assistance most heavily.3,8 Not surprisingly, Kentucky residents were more likely to receive application assistance, given the state’s aggressive outreach efforts. 9 However, navigators can only be effective after applicants seek assistance, and thus must already be at least somewhat knowledgeable about the ACA.
Our study has several limitations. We can only present correlations between our covariates and outcome variables, as opposed to causal relationships. Furthermore, our sample’s racial composition differs significantly between states: More than 80% of the Latinos in our sample live in Texas, whereas other racial and ethnic groups were more evenly distributed, though we adjusted directly for state of residence in our analysis. Our study also focused exclusively on those reporting US citizenship and did not ask about the legal status of other household members. This means our results are not due to differences in eligibility for public coverage due to immigration status among survey respondents. However, having family members with undocumented status can reduce coverage rates among legal immigrants and US citizens, and our data set did not allow us to explore this important issue. Finally, our response rate of 26% may be subject to non-response bias, though we weighted to appropriate population benchmarks, which research shows can mitigate this potential bias. 10
In conclusion, despite abundant research pointing to large drops in uninsured rates associated with the ACA and Medicaid expansion,2,11-13 disparities remain for Latinos. Our results highlight the need to target outreach efforts to low-income adults with less education—a group with a disproportionately high number of Latinos—to close gaps in ACA awareness and, ultimately, health insurance coverage rates.
Footnotes
Authors’ Note
Dr. Sommers currently serves part-time as an advisor in the Office of the Assistant Secretary for Planning and Evaluation, at the U.S. Department of Health and Human Services.The views presented here are those of the authors and do not represent U.S. Department of Health and Human Services, Agency for Healthcare Research and Quality, or the Commonwealth Fund.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was supported by a research grant from the Commonwealth Fund. Dr. Sommers’ work on this project was also supported in part by a grant (K02HS021291) from the Agency for Healthcare Research and Quality.