Medicaid Expansion Increased Coverage, Improved Affordability, And Reduced Psychological Distress For Low-Income Parents

When the Affordable Care Act (ACA) was passed in 2010, it included a Medicaid expansion that aimed to reduce uninsurance among adults with incomes at or below 138 percent of the federal poverty level. However, the June 2012 Supreme Court ruling in National Federation of Independent Business v. Sebelius made that expansion optional. As of April 2017, thirty-two states (including the District of Columbia) had chosen to expand eligibility for Medicaid under the ACA. Much attention has been focused on potential coverage gains for childless adults, ¹ but Medicaid eligibility for parents was also limited in many states before passage of the ACA. Several states, including Alabama, Arkansas, Louisiana, Missouri, and Texas, had income eligibility thresholds below 30 percent of poverty. As a result of this limited Medicaid eligibility in some states as well as other factors, over ten million parents were uninsured in 2010. ² Studies have found significant effects of previous Medicaid expansions on parents’ coverage, access to care, use of services, and mental health status, ^3–5 and those findings suggest that uninsured parents are likely to benefit from the ACA Medicaid expansions as well. Furthermore, children may experience positive spillover effects from their parents’ gains in coverage and improved access to care, health, and financial well-being. ^6,7

Parents with incomes at or below 138 percent of poverty, but above their state’s pre-ACA Medicaid eligibility threshold, became newly eligible for Medicaid coverage in states that opted to participate in the ACA Medicaid expansion. Coverage options remained limited for low-income parents in nonexpansion states, where Medicaid income eligibility thresholds were often well below the poverty level (with an average threshold of 52 percent of poverty). However, in nonexpansion states, parents with incomes of 100–138 percent of poverty were eligible for federal subsidies to purchase coverage in the health insurance Marketplaces, and in both expansion and nonexpansion states parents who were already eligible for Medicaid may have had an increased probability of enrollment after 2014 as a result of publicity, outreach, and enrollment efforts associated with the ACA expansions.

Strong and consistent evidence has emerged that the ACA Medicaid expansions increased Medicaid coverage and reduced uninsurance rates for low-income adults. ^8–10 Several studies have also found increases in access to care and service use and reductions in out-of-pocket spending. ^9,11,12 Descriptive evidence has shown coverage and access improvements for all parents under the ACA. ¹³ However, findings on the impacts of the Medicaid expansion for parents have been mixed. One study found no significant increases in insurance coverage or access to care for low-income parents through 2015, ¹⁴ while another study found that the Medicaid expansion reduced uninsurance rates for parents with a high school education or less in 2014. ¹⁵ Neither of these studies accounted for the variation in the magnitude of the expansion for parents across states that resulted from states’ very different Medicaid eligibility thresholds before implementation of the ACA Medicaid expansion.

In this article we describe changes in insurance coverage, access to care, service use, affordability of care, and health status for low-income parents through 2015. We attempt to isolate the impacts of the ACA Medicaid expansion using variations in the Medicaid eligibility threshold for parents within states over time. We also describe the missed opportunities for parents in states that did not expand Medicaid under the ACA.

Study Data And Methods

Data And Sample

We used public use data for the period 2010–15 from the Integrated Health Interview Series of the National Health Interview Survey (NHIS). ¹⁶ These data provide harmonized versions of NHIS variables across data years. We also obtained access to state and county identifiers through the Research Data Center of the National Center for Health Statistics.

We defined parents as US citizen adults ages 19–64 with a biological, step-, or adopted child ages 0–18 years in their health insurance unit—that is, a group of family members who would be considered a family in determining eligibility for Marketplace subsidies, Medicaid, or family/dependent coverage through private insurance (this is a more narrow definition of family than that used for other purposes). We also constructed a measure of income relative to poverty for the health insurance unit using NHIS earnings and income information, which includes imputed information for approximately 25 percent of the sample, and poverty guidelines from the Department of Health and Human Services.

Our sample included parents whose health insurance unit had an income at or below 138 percent of poverty. We excluded noncitizens because legal residents face additional restrictions on Medicaid eligibility, and undocumented immigrants are not eligible for Medicaid. We also excluded people who were pregnant or covered by Medicare at the time of the survey and those who had received Supplemental Security Income benefits in the previous calendar year, because those groups are subject to different Medicaid eligibility criteria.

We constructed a health insurance hierarchy that had seven mutually exclusive categories: Medicare, Marketplace insurance, insurance sponsored by the employer (including the military), other private insurance, Medicaid or the Children’s Health Insurance Program (CHIP), other public insurance, and no insurance. We report data for four categories: employer-sponsored insurance; Medicaid or CHIP; Marketplace or other private or other public insurance; and no insurance. These categories are based on self-reported information and reflect coverage at the time of survey. We also examined whether parents reported that their health insurance was better, worse, or the same, compared to the previous year.

We constructed several measures of access and use: the percentages of parents who had a usual source of care (other than the emergency department [ED]) and who in the past twelve months had seen a general doctor or any provider (a general doctor, specialist, obstetrician/gynecologist, midlevel provider, or mental health provider). We also identified parents who in the past year had had multiple ED visits or trouble finding a provider who would see them, and those who delayed care for noncost reasons (issues with transportation, wait times for appointments or in the provider’s office, inconvenient office hours, or trouble getting through on the phone).

We measured affordability of care by identifying parents who reported being very worried, moderately worried, or not worried at all about paying either for the medical costs of a serious illness or injury or for costs of routine health care. We also measured the percentages of parents who reported in the past year having had trouble paying medical bills for themselves or their family members; not receiving needed medical care, prescription drugs, or mental health care because of cost; and having delayed care because of cost.

Finally, we measured self-reported general health status (excellent or very good, good, or fair or poor) at the time of the survey and mental health status in the previous thirty days using the Kessler K6 Psychological Distress Scale. ¹⁷ We classified respondents into three categories of psychological distress: none or mild (with a score on the scale of 0–7), moderate (8–12), or severe (13 or more). We also examined whether parents reported that their health was better, worse, or the same, compared to the previous year. Additional details on sample sizes and variable construction are available in the online Appendix. ¹⁸

Methods

We chose the outcomes described above based on the Andersen model of access to care ¹⁹ and previous work on Medicaid expansions. ^20–22 Expanding Medicaid eligibility is expected to increase health insurance coverage, which has the potential to strengthen patients’ access to affordable services, enhance their financial security, and ultimately improve their general and mental health status. Medicaid expansions can also crowd out employer-sponsored coverage, however, and improvements in access, affordability, and health outcomes depend on enrollees’ ability to navigate the health care system and that system’s capacity to meet increased demand for care.

While we generally hypothesized that there would be increases in coverage and improvements in access, affordability, and health status under the Medicaid expansion, the expected effects on ED use, trouble finding a provider, and delayed care for noncost reasons were less clear. Medicaid expansion could reduce ED use if new enrollees gained access to office-based providers, but it could also increase ED use if the cost of that use declined for the newly insured. Similarly, new enrollees might have less trouble finding a provider after gaining coverage, but if capacity is an issue, they might instead have more trouble finding a provider.

We first estimated changes in insurance coverage, access and use, affordability, and health status for low-income parents before and after the 2014 Medicaid expansions, separately for expansion and nonexpansion states. We classified as expansion states those twenty-six states (including the District of Columbia) that had expanded Medicaid by April 2014. We excluded Indiana, New Hampshire, and Pennsylvania—which expanded Medicaid in late 2014 or early 2015—from our main analyses so that we could focus on the effects of the 2014 expansions over two years.

We also estimated the simple (unadjusted) difference-in-differences between expansion and nonexpansion states over time to begin to isolate the effects of the Medicaid expansion from other changes occurring in the study period. However, these estimates did not account for the variation in the magnitude of the expansions across states for parents or for other differences between the populations in expansion and nonexpansion states.

To address these issues, we used a multivariate difference-in-differences approach with a continuous policy variable that reflected the Medicaid eligibility threshold for parents in a given state and year. Specifically, we estimated a model with state and year fixed effects to exploit the variation in the Medicaid eligibility threshold within states over time. To increase the precision of our estimates, we pooled NHIS data in two-year intervals (2010–11, 2012–13, and 2014–15) and assigned individuals their state Medicaid eligibility threshold for the earlier year in each pair. This approach allowed us to capture an average effect of the 2014 Medicaid expansions in 2014 and 2015.

We compiled information on state Medicaid eligibility rules for working parents in 2010, 2012, and 2014 from the Henry J. Kaiser Family Foundation (complete citations are in the Appendix). ¹⁸ On average, the Medicaid eligibility threshold for parents in expansion states increased from 112 percent of poverty in 2012 to 146 percent in 2014, but this average obscures substantial variation across states (Appendix Table 2). ¹⁸ The largest eligibility expansions occurred in Arkansas, West Virginia, and Oregon—which had increases in the eligibility threshold of 122, 107, and 99 percentage points, respectively. Importantly, the expansion states had much higher eligibility thresholds in 2012 than the nonexpansion states did (112 percent versus 60 percent), so the potential gains for nonexpansion states were considerably larger, on average, than the actual gains in the participating states. In addition, six expansion states and two nonexpansion states had expanded eligibility to parents beyond the ACA threshold of 138 percent of poverty before 2014. We top-coded the eligibility threshold at 138 percent in all analyses because changes at higher income thresholds were unlikely to affect our sample of low-income parents. Additional details on the eligibility rules, including our use of working rather than jobless parent thresholds and the implications for our analysis, are available in the Appendix. ¹⁸

For ease of interpretation, we estimated linear probability models on binary measures of coverage, access and use, affordability, and health status, and we included parent-level controls for age, sex, race/ethnicity, education, work status, income as a percentage of poverty, marital status, number of children, and presence of an activity limitation. To further account for changing economic conditions, we also controlled for the county employment rate. We clustered standard errors at the state level and adjusted them to account for the multiple imputations of income. Additional details and means for all covariates are available in Appendix Table 3. ¹⁸ Our key variable of interest was the state Medicaid income eligibility threshold for parents, measured as a percentage of poverty, and the coefficient of interest reflected the effect of a 100-percentage-point increase in the eligibility threshold on the outcome of interest.

The difference-in-differences approach relies on the assumption that preexisting trends in the outcomes of interest are similar in treatment and comparison groups. In our case, with a continuous policy variable, it required that the preexisting trends not be correlated with changes in the eligibility threshold. To test this assumption, we estimated a model that included state-specific linear trends in addition to state and year fixed effects. We were unable to estimate this model for our measures of worries about medical care costs or psychological distress because we had only one year of preexpansion data for these measures.

Given the variety of methodological approaches available to estimate the impacts of Medicaid eligibility expansions, we also tested the sensitivity of our results by using the simulated eligibility approach pioneered by Janet Currie and Jonathan Gruber. ²³ We imputed individual eligibility to our sample of low-income parents based on state, year, and income. We then drew a national sample of 3,000 parents and applied the eligibility rules for each state and year to the sample to generate the simulated eligibility instrument, or the share of the national sample that would be eligible under each state’s rules. We estimated the model using two-stage least squares, with the endogenous eligibility indicator as our main variable of interest. We provide additional details on the advantages and disadvantages of each approach and discuss other robustness checks in the Appendix. ¹⁸

To investigate nonlinearities in the relationship between the eligibility threshold and our outcomes of interest, we replaced the continuous threshold with four categorical variables: income eligibility thresholds of less than 50 percent of poverty, of 50–99 percent of poverty, of 100–137 percent of poverty, and of 138 percent of poverty and above. We then estimated the effect of being in a state with an eligibility threshold of 138 percent of poverty and above compared to each of the other categories, to capture the separate effects of small, medium, and large eligibility expansions under the ACA.

Finally, we used the results of our threshold model to predict the insurance coverage status of low-income parents in nonexpansion states if the eligibility threshold in their state had increased to 138 percent of poverty. This approach assumed that people in nonexpansion states would respond to a Medicaid expansion as similar people in expansion states did.

Limitations

This analysis had several limitations. First, there is measurement error in the eligibility thresholds, incomes, and types of health insurance coverage. Specifically, we could not reliably determine the appropriate threshold for an individual parent based on the NHIS data on work status, so we used the threshold for working parents in our main specification and tested the sensitivity of our results to using the nonworking threshold. In addition, we allocated income across the health insurance units that made up a family, and NHIS income measures refer to annual income in the previous calendar year (for example, income reported in the 2014 survey refers to 2013 annual income). As a result, using our income measure to approximate the Medicaid target population was subject to error. Furthermore, reports of the presence or absence of coverage are generally valid, but measurement error is more likely in reports of the type of coverage and is likely to be increasing with the changes introduced under the ACA. ²⁴

Second, we analyzed two measures that captured perceptions of coverage and health status compared to the previous year. With respect to coverage, we would expect any reported improvement to occur immediately after a respondent gained coverage in either 2014 or 2015. With respect to health status, the likely timing and persistence of any improvements are ambiguous. Thus, pooling 2014 and 2015 data for health insurance compared to the previous year might understate reported coverage improvements if coverage gains were concentrated in 2014. Any likely bias in pooling data on health status compared to the previous year would be less obvious, but our estimates reflect an average of reported changes in 2014 and 2015. Furthermore, all of our outcome measures were self-reported and could be subject to recall or social desirability bias.

Third, there could be unobserved factors at the individual or state level that were correlated with the magnitude of the eligibility expansions and the outcomes of interest. For example, if states with larger expansions invested more resources in outreach and education, compared to states with smaller expansions, or if parents in states with larger expansions differed from those in states with smaller expansions on characteristics not captured in the regression analysis (such as severity of health care need), the estimates of differences in outcomes by the size of the expansion would also reflect the effects of these other factors.

Fourth, relatively small sample sizes for some analyses reduced our ability to detect small changes, and the design of the NHIS makes it likely that we underestimated the full effects of the expansion at two years—given the continuous fielding of the NHIS over a given year and the need to rely on many survey questions that are based on experiences during the previous twelve months.

Finally, we designed this analysis to detect the overall effect of the eligibility expansion on the outcomes of interest, not the effects of gaining Medicaid coverage on access or affordability or on health status. Thus, our ability to detect these second-order effects was more limited than our ability to detect effects on insurance coverage.

Study Results

All results reported in the text are significant at the 5 percent level ( $p<0.05$ ) unless otherwisenoted.

Changes In Expansion And Nonexpansion States

Based on simple comparisons over time in both expansion and nonexpansion states, we found that insurance coverage for low-income parents changed significantly after the ACA’s 2014 Medicaid expansions. The uninsurance rate for parents in expansion states fell 13.0 percentage points from 2012–13 to 2014–15, and there was a nearly corresponding increase in Medicaid or CHIP coverage ( Exhibit 1 ). The share of parents in expansion states who reported that their coverage was better than in the previous year also increased ( $p<0.10$ ). In nonexpansion states, the uninsurance rate fell by 10.6 percentage points, a change driven by increases in Medicaid or CHIP (4.0 percentage points, $p>0.10$ ), employer-sponsored coverage (3.0 percentage points, $p<0.10$ ), and other coverage (3.6 percentage points). When we compared the unadjusted changes in coverage in expansion and nonexpansion states over time, however, only the unadjusted difference-in-differences for employer-sponsored coverage was marginally significant ( $p<0.10$ ).

We also found significant increases in access and use among low-income parents in expansion states. The share of parents in those states who had a usual source of care and who had had a general doctor visit or any provider visit increased ( Exhibit 1 ). There were also strong improvements in almost every affordability measure examined for parents in expansion states. Changes in health status were mixed, with a decline in the shares of parents who reported that their health was better than in the previous year and who reported severe psychological distress following the expansions.

There were no significant changes in access and use or health status in nonexpansion states, but there were strong improvements in several affordability measures. When we compared the unadjusted changes in access and use, affordability, and health status in expansion and nonexpansion states, only the unadjusted difference-in-differences on delayed care because of noncost reasons was marginally significant ( $p<0.10$ ). This finding suggests that there was an increase in non-cost-related delays among parents in expansion states relative to nonexpansion states.

Impact Estimates Accounting For The Size of Medicaid Eligibility Expansions For Low-Income Parents

To better isolate the impacts of the Medicaid expansion on low-income parents, we estimated multivariate models that accounted for the characteristics of the parents and the variation in the magnitude of the expansions to parents across states. We found that a 100-percentage-point increase in the Medicaid income eligibility threshold would result in an 11.0-percentage-point decrease in uninsurance and a 14.6-percentage-point increase in Medicaid or CHIP coverage, if all else were equal ( Exhibit 2 ). The estimated effect on employer-sponsored coverage was a decline of 5.2 percentage points ( $p<0.10$ ), which suggests some evidence of crowd-out of employer-sponsored coverage. These estimates suggest that the average change in income eligibility thresholds in expansion states of 34 percentage points (Appendix Table 2) ¹⁸ decreased uninsurance rates by about 3.7 percentage points. Consistent with these coverage gains, an increase in the eligibility threshold also increased the share of parents reporting better coverage than in the previous year.

We found no significant overall improvements in access and use for parents in response to an increase in the Medicaid eligibility threshold. Increasing the threshold reduced problems in paying family medical bills, but we found no other significant effects on affordability. Finally, increasing the threshold reduced the share of low-income parents who reported severe psychological distress ( $p<0.10$ ) and increased the share who reported no or mild psychological distress.

When we added controls for state linear trends, we generally found similar results with reduced precision. However, we also found a marginally significant decline in delayed care because of cost. Using the simulated eligibility approach also resulted in findings that were very similar to those in our main specification.

Our investigation of nonlinearities in the relationship between the Medicaid income eligibility threshold and our outcomes revealed some interesting patterns. The estimates can be interpreted as the effect of moving from a state with one of the lower thresholds to a state with eligibility of at least 138 percent of poverty. These estimates thereby capture the separate effects of small (from 100–137 percent of poverty), medium (from 50–99 percent of poverty) and large (from less than 50 percent of poverty) eligibility expansions to 138 percent of poverty.

We found that expansions of all sizes had significant effects on rates of uninsurance and Medicaid/CHIP coverage and that the magnitude of the effects increased as the size of the expansion did ( Exhibit 3 ). These findings support the assumption of linearity in our main specification. Similarly, the patterns for quality of insurance coverage compared to the previous year and having problems paying medical bills support the findings from our main model. On measures of access and use, however, we found that small expansions were associated with an increased probability of having a usual source of care ( $p<0.10$ ), and having had a general doctor visit and any provider visit, compared to the larger expansions. Small expansions were also associated with an increase in having trouble finding a provider.

We also found evidence of reductions in unmet needs and delayed care because of cost that resulted from medium-size expansions. Finally, we found similar reductions in severe psychological distress associated with large and medium-size expansions, but large expansions were associated with a shift toward no or mild psychological distress, while medium expansions were associated with a shift toward moderate distress. Altogether, this analysis suggests that our main specification generally captured the effects of the Medicaid expansion on coverage, affordability, and psychological distress but did not capture the effects of small expansions on access and use.

We explored a variety of additional subgroup analyses and robustness checks on our main specification. For example, we found that men and women experienced similar coverage changes in response to the Medicaid expansion, but women had an increase in doctor visits and a reduction in worries about costs, while reductions in psychological distress were concentrated among men (Appendix Table 4). ¹⁸ We also found results that were generally consistent, but smaller in magnitude, when we included noncitizens in the sample (Appendix Table 5). ¹⁸ And we found additional evidence of reduced affordability problems when we used the Medicaid income eligibility threshold for nonworking parents (Appendix Table 6). ¹⁸ These and other sensitivity analyses are discussed in more detail in the Appendix. ¹⁸

As indicated above, the states that expanded Medicaid under the ACA already had much higher eligibility thresholds for parents, compared to the states that did not expand. Nonexpansion states would have experienced, on average, a 78-percentage-point increase in their Medicaid eligibility threshold for parents if they had opted to expand eligibility (Appendix Table 2). ¹⁸ If the nonexpansion states had expanded Medicaid in 2014, our model suggests that the uninsurance rate among low-income parents would have fallen to an average rate of 24.3 percent in 2014–15, compared to the actual 2014–15 uninsurance rate of 33.7 percent ( Exhibit 4 ). We estimate that the Medicaid/CHIP coverage rate in 2014–15 would have increased to 47.0 percent in nonexpansion states, compared to the actual 2014–15 rate of 32.7 percent. This would have been offset by an estimated decline in employer-sponsored coverage that was not significant.

SOURCE Authors’ analysis of data for 2010–15 from the National Health Interview Survey. NOTES Low-income parents, the sample, nonexpansion states, and “other coverage” are explained in the Notes to Exhibit 1 . Significance refers to the difference from the actual percentage. CHIP is Children’s Health Insurance Program. ** $p<0.05$

Discussion

We estimated the effects of the ACA Medicaid expansion on insurance coverage, access to care, service use, affordability of care, and health status for low-income parents. In contrast to previous studies of the ACA expansion, ^9–12 we accounted for the wide variation in the size of the expansion for parents across states to better capture the average impact on parents, and we specifically estimated the effects of expansions of different sizes.

We found strong and consistent evidence that the Medicaid expansion increased Medicaid coverage and reduced uninsurance rates among low-income parents in 2014–15. We also found some evidence of a reduction in rates of employer-sponsored coverage, but this result was more sensitive to the model specification and disappeared when we focused on parents with incomes below poverty (Appendix Table 6). ¹⁸ Our results suggest that low-income parents in nonexpansion states would have experienced an additional 9.4-percentage-point drop in their uninsurance rate—a decline of 28 percent—if those states had opted to participate in the ACA Medicaid expansion.

We found that only smaller Medicaid expansions were associated with an increased probability of having a visit with a general doctor or other provider in the previous year, compared to larger expansions. However, smaller expansions had no effects on affordability of care. In contrast, we found that both medium-size and large expansions reduced problems paying medical bills and that medium-size expansions also reduced delayed care and unmet need because of cost.

It is important to remember that the size of the expansion is explicitly tied to the income of the target population. Thus, small expansions target parents with somewhat higher incomes, while large expansions target a broader group—including those with very low incomes. This suggests that expansions in different states reached parents with different characteristics (for example, varying degrees of financial resources and health needs). While we included some controls for these characteristics, there might still be unobserved factors that could have contributed to our results.

Finally, we found a meaningful impact of the Medicaid expansion on mental health for low-income parents, with significant reductions in severe psychological distress concentrated among states with medium and large expansions. Given the lack of impacts on service use and the significant improvements in affordability of care in these states, the findings on psychological distress could suggest that the security of having health insurance provides mental health benefits beyond those obtained through medical care. This is consistent with evidence on the “warm glow” of health insurance from the Oregon Health Insurance Experiment. ²²

Conclusion

We found strong and consistent evidence that the ACA Medicaid expansion increased coverage, reduced problems paying medical bills, and reduced psychological distress among low-income parents. We also found important missed opportunities for coverage gains among nonexpansion states. Importantly, this analysis might underestimate the potential gains for nonexpansion states because we had limited power to detect the effects of large expansions. Only three expansion states had eligibility levels below 50 percent of poverty before implementation of the ACA Medicaid expansion, but thirteen nonexpansion states had thresholds that low (Appendix Table 2). ¹⁸

The benefits of the Medicaid expansion to low-income parents also have the potential to produce spillover effects for low-income children. Evidence suggests that children benefit when their parents are insured, and the mental health improvements for parents gaining coverage under the ACA could have particularly strong effects on the health and well-being of their children. As policy makers continue to debate the future of the ACA, this study provides important evidence on the benefits of expanding Medicaid eligibility for low-income parents and the missed opportunities for states not participating in the ACA Medicaid expansion.

ACKNOWLEDGMENTS

This work was funded by the Robert Wood Johnson Foundation. The authors are grateful to Linda Blumberg and John Holahan for comments on an earlier version, and to Patricia Barnes and the staff at the Research Data Center of the National Center for Health Statistics for their help with this study. The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the Research Data Center; the National Center for Health Statistics; the Centers for Disease Control and Prevention; or the Urban Institute, its trustees, or its funders.

NOTES

• 1 Sommers BD , Tomasi MR , Swartz K , Epstein AM . Reasons for the wide variation in Medicaid participation rates among states hold lessons for coverage expansion in 2014 . Health Aff (Millwood) . 2012 ; 31 ( 5 ): 909 – 19 . Go to the article, Google Scholar
• 2 Heberlein M , Huntress M , Kenney G , Alker J , Lynch V , Mancini T . Medicaid coverage for parents under the Affordable Care Act [Internet]. Washington (DC) : Georgetown University Health Policy Institute ; 2012 Jun [cited 2017 Mar 27 ]. Available from: http://ccf.georgetown.edu/wp-content/uploads/2012/08/Medicaid-Coverage-for-Parents.pdf Google Scholar
• 3 McMorrow S , Kenney GM , Long SK , Goin DE . Medicaid expansions from 1997 to 2009 increased coverage and improved access and mental health outcomes for low-income parents . Health Serv Res . 2016 ; 51 ( 4 ): 1347 – 67 . Crossref, Medline, Google Scholar
• 4 Hamersma S , Kim M . Participation and crowd out: assessing the effects of parental Medicaid expansions . J Health Econ . 2013 ; 32 ( 1 ): 160 – 71 . Crossref, Medline, Google Scholar
• 5 Busch SH , Duchovny N . Family coverage expansions: impact on insurance coverage and health care utilization of parents . J Health Econ . 2005 ; 24 ( 5 ): 876 – 90 . Crossref, Medline, Google Scholar
• 6 Dubay L , Kenney G . Expanding public health insurance to parents: effects on children’s coverage under Medicaid . Health Serv Res . 2003 ; 38 ( 5 ): 1283 – 301 . Crossref, Medline, Google Scholar
• 7 Davidoff A , Dubay L , Kenney G , Yemane A . The effect of parents’ insurance coverage on access to care for low-income children . Inquiry . 2003 ; 40 ( 3 ): 254 – 68 . Crossref, Medline, Google Scholar
• 8 Antonisse L , Garfield R , Rudowitz R , Artiga S . The effects of Medicaid expansion under the ACA: findings from a literature review [Internet]. Menlo Park (CA) : Henry J. Kaiser Family Foundation ; 2016 Jun [cited 2017 Mar 27 ]. Available from: http://kff.org/medicaid/issue-brief/the-effects-of-medicaid-expansion-under-the-aca-findings-from-a-literature-review/ Google Scholar
• 9 Wherry LR , Miller S . Early coverage, access, utilization, and health effects associated with the Affordable Care Act Medicaid expansions: a quasi-experimental study . Ann Intern Med . 2016 ; 164 ( 12 ): 795 – 803 . Crossref, Medline, Google Scholar
• 10 Courtemanche C , Marton J , Ukert B , Yelowitz A , Zapata D . Early impacts of the Affordable Care Act on health insurance coverage in Medicaid expansion and non-expansion states . J Policy Anal Manage . 2016 ; 36 ( 1 ): 178 – 210 . Crossref, Google Scholar
• 11 Sommers BD , Gunja MZ , Finegold K , Musco T . Changes in self-reported insurance coverage, access to care, and health under the Affordable Care Act . JAMA . 2015 ; 314 ( 4 ): 366 – 74 . Crossref, Medline, Google Scholar
• 12 Golberstein E , Gonzales G , Sommers BD . California’s early ACA expansion increased coverage and reduced out-of-pocket spending for the state’s low-income population . Health Aff (Millwood) . 2015 ; 34 ( 10 ): 1688 – 94 . Go to the article, Google Scholar
• 13 Karpman M , Kenney GM , McMorrow S , Gates JA . Health care coverage, access, and affordability for children and parents: new estimates for March 2016 [Internet]. Washington (DC) : Urban Institute , 2016 Sep 14 [cited 2017 Mar 27 ]. Available from: http://hrms.urban.org/briefs/coverage-access-affordability-children-parents.pdf Google Scholar
• 14 Simon K , Soni A , Cawley J . The impact of health insurance on preventive care and health behaviors: evidence from the 2014 ACA Medicaid expansions [Internet]. Cambridge (MA) : National Bureau of Economic Research ; 2016 May [cited 2017 Mar 27 ]. (NBER Working Paper No. 22265). Available for download (fee required) from: http://www.nber.org/papers/w22265 Google Scholar
• 15 Kaestner R , Garrett B , Gangopadhyaya A , Fleming C . Effects of ACA Medicaid expansions on health insurance coverage and labor supply [Internet]. Cambridge (MA) : National Bureau of Economic Research ; 2015 Dec [cited 2017 Mar 27 ]. (NBER Working Paper No. 21836). Available for download (fee required) from: http://www.nber.org/papers/w21836 Google Scholar
• 16 IPUMS Health Surveys . NHIS [Internet]. Minneapolis (MN) : University of Minnesota ; [cited 2017 Apr 5 ]. Available from: https://ihis.ipums.org/ihis/ Google Scholar
• 17 Kessler RC , Barker PR , Colpe LJ , Epstein JF , Gfroerer JC , Hiripi E , et al. Screening for serious mental illness in the general population . Arch Gen Psychiatry . 2003 ; 60 ( 2 ): 184 – 9 . Crossref, Medline, Google Scholar
• 18 To access the Appendix, click on the Appendix link in the box to the right of the article online.
• 19 Andersen RM . Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav . 1995 ; 36 ( 1 ): 1 – 10 . Crossref, Medline, Google Scholar
• 20 Institute of Medicine . America’s uninsured crisis: consequences for health and health care . Washington (DC) : National Academies Press , 2009 . Google Scholar
• 21 Gross T , Notowidigdo MJ . Health insurance and the consumer bankruptcy decision: evidence from the expansion of Medicaid . J Public Econ . 2011 ; 95 ( 7–8 ): 767 – 78 . Crossref, Google Scholar
• 22 Finkelstein A , Taubman S , Wright B , Bernstein M , Gruber J , Newhouse JP , et al. The Oregon Health Insurance Experiment: evidence from the first year . Q J Econ . 2012 ; 127 ( 3 ): 1057 – 106 . Crossref, Medline, Google Scholar
• 23 Currie J , Gruber J . Health insurance eligibility, utilization of medical care, and child health . Q J Econ . 1996 ; 111 ( 2 ): 431 – 66 . Crossref, Google Scholar
• 24 Pascale J . Measurement error in health insurance reporting . Inquiry . 2008–2009 ; 45 ( 4 ): 422 – 37 . Crossref, Medline, Google Scholar