Abstract
Introduction
We use data from the Behavioral Risk Factor Surveillance System (BRFSS) from 2012 to 2015 to estimate the effects of the Affordable Care Act's (ACA) Medicaid expansions on insurance coverage and access to care for low-income women of reproductive age (19–44).
Methods
We use two-way fixed effects difference-in-differences models to estimate the effects of Medicaid expansions on low-income (<100% of the Federal Poverty Level) women of reproductive age. Additional models are stratified to estimate effects based on women's parental status, pre-ACA state Medicaid eligibility levels, and the presence of a state Medicaid family planning waiver.
Results
ACA Medicaid expansions decreased uninsurance among low-income women of reproductive age by 13.2 percentage points. This decrease was driven by a decrease of 27.4 percentage points for women without dependent children, who also experienced a decrease in the likelihood of not having a personal doctor (13.3 percentage points). We find a 3.8-percentage point reduction in the likelihood of experiencing a cost barrier to care among all women, but no significant effects for other access measures or subgroups. When stratified by state policies, decreases in uninsurance were greater in states expanding from pre-ACA eligibility levels of less than 50% of Federal Poverty Level (19.4 percentage points) and in states without a Medicaid family planning waiver (17.6 percentage points).
Conclusions
The ACA Medicaid expansion increased insurance coverage for low-income women of reproductive age, with the greatest effects for women without dependent children and women residing in states with relatively lower pre-ACA Medicaid eligibility levels or with no family planning waiver before the ACA.
The 2010 Affordable Care Act (ACA) contains several provisions intended to increase health insurance coverage, including the 2014 Medicaid eligibility expansion to all nonelderly adults with family incomes of up to 138% of the Federal Poverty Level (FPL). This provision was originally intended to occur across all states, but a 2012 Supreme Court decision made state Medicaid expansions optional. This choice, given to the states, created a “natural experiment,” which allows researchers to use the variation in expansion decisions across states to measure the effects of ACA Medicaid expansions. As of January 2017, 31 states and the District of Columbia expanded Medicaid eligibility under the ACA provision, some using waivers to customize their expansion (
Kaiser Family Foundation, 2017
). Although there are reported promising effects of ACA Medicaid expansion on insurance coverage, access to care, and use of preventive services for low-income individuals overall, to date, no research has studied the effects of this policy for women of reproductive age (Antonisse et al., 2016
, Courtemanche et al., 2017
; McMorrow et al., 2017
, Miller and Wherry, 2017
, Simon et al., 2017
, Sommers et al., 2015
, Wherry and Miller, 2016
).Health insurance can provide reproductive-age women with improved access to medical care by reducing the patient and clinician financial burden associated with obtaining needed services, particularly for low-income women (
Biermann et al., 2006
, Jack and Culpepper, 1990
, Johnson et al., 2006
, - Johnson K.
- Posner S.F.
- Biermann J.
- Codero J.F.
- Atrash H.K.
- Parker C.S.
- Curtis M.G.
Recommendations to improve preconception health and health care – United States: A Report of the CDC/ATSDR Preconception Care Work Group and the Select Panel on Preconception Care (pp. 1–23.
Centers for Disease Control and Prevention,
Atlanta, GA2006
Johnson et al., 2015
, Lu et al., 2006
, Lu and Halfon, 2003
, Salganicoff and An, 2008a
). Beyond women's own health, increasing access to medical care for nonpregnant women can also have positive effects for any future children, as women's access to and use of health care outside of pregnancy has been identified as necessary to improving both maternal and child health outcomes (Atrash et al., 2006
, Johnson et al., 2015
, Lu and Halfon, 2003
, McCormick, 2001
).Background
Historically, the eligibility structure of Medicaid focused on providing coverage to low-income pregnant women and mothers, largely omitting nonpregnant and childless women (
Adams and Johnston, 2016
, Salganicoff and An, 2008b
). Recognizing the importance of prenatal care to promoting healthy pregnancies and healthy children, the federal government mandated that states expand Medicaid eligibility to pregnant women with incomes up to 133% of the FPL in the late 1980s, and most states further expanded eligibility to those with higher incomes (Johnson et al., 2015
, Kaiser Family Foundation, 2016
). However, this pregnancy-based coverage only covers women during pregnancy and immediately postpartum (usually 60 days), leaving many women of reproductive age uninsured before and after pregnancy (MACPAC., 2014
). Before the ACA, only 16 states and the District of Columbia covered working parents with family with incomes of up to 100% of the FPL or more, and only five states provided Medicaid coverage for childless adults (Artiga et al., 2010
, Ross et al., 2009
).- Ross D.C.
- Jarlenski M.
- Artiga S.
- Marks C.
A foundation for health reform: Findings of an annual 50-state survey of eligibility rules, enrollment and renewal procedures and cost-sharing practices in Medicaid and CHIP for children and parents during 2009 (Kaiser Commission on Medicaid and the Uninsured No. 8028).
Kaiser Family Foundation,
Washington, DC2009
Thus, pre-ACA Medicaid eligibility for low-income women depended not only on their income, but also on their pregnancy status, parental status, disability status, and their state of residence. Some states partially addressed this gap in coverage by implementing Medicaid family planning waivers and state plan amendments. These programs expanded eligibility for reproductive health and family planning services to low-income women who did not otherwise qualify for the Medicaid program. Although these programs do not provide women with comprehensive health insurance coverage (
Atrash et al., 2006
, Johnson et al., 2015
, Lu and Halfon, 2003
, McCormick, 2001
, Sonfield et al., 2011
), it is unclear if women consider this type of coverage to include access to primary care services.Earlier Studies
Expansions of Medicaid eligibility to women of reproductive age before the ACA were found to increase insurance coverage, use of prenatal care services and cervical cancer screens, and timely initiation of prenatal care among low-income pregnant women (
Currie and Gruber, 1996
, Dubay et al., 2001
, Howell, 2001
, Sabik et al., 2017
). A number of recent papers have studied the effects of the ACA Medicaid expansion on low-income individuals, finding that expansions increased coverage through Medicaid programs (Antonisse et al., 2016
, Courtemanche et al., 2017
, McMorrow et al., 2017
, Miller and Wherry, 2017
, Simon et al., 2017
, Sommers et al., 2015
, Wherry and Miller, 2016
). Expansions were also found to increase access to care by decreasing the likelihood that individuals do not have a personal physician, have difficulty accessing medicine, are unable to afford needed follow-up care, and worry about paying medical bills (Miller and Wherry, 2017
, Simon et al., 2017
, Sommers et al., 2015
). The effect of expansions on health care use is mixed; some studies found increases in physician and hospital visits, whereas others found no effect on use (Miller and Wherry, 2017
, Wherry and Miller, 2016
). Finally, ACA Medicaid expansions were found to increase the use of certain forms of preventive care (Simon et al., 2017
).Among these studies, only two specifically report effects for women.
Courtemanche et al., 2017
find that ACA Medicaid expansions increased insurance coverage among all women by 3.4 percentage points. Simon et al., 2017
report effects for women and for childless adults versus parents, limiting all samples to individuals with an income of less than 100% of the FPL. They find that expansions increased insurance coverage for women by 3.4 percentage points and for childless adults by 10.1 percentage points, with no significant increase for parents. They also find greater effects in the 21 states that had no Medicaid eligibility for childless adults before 2014, states in which expansion increased eligibility from none to 138% of the FPL. Their analysis excludes the 10 states that covered some of this population before the ACA and, therefore, experienced smaller changes in eligibility levels. For example, Delaware covered childless adults up to 100% of the FPL in its Diamond State Health Plan waiver before the ACA. Thus, the expansion in Delaware only changed eligibility from 100% to 138% of the FPL. These results suggest that eligibility gains among women of reproductive age may vary based on women's parental status and state pre-ACA eligibility policies, but further research is needed to study these questions specifically for women of reproductive age.Contribution
In this study, we build on prior research by investigating the effects of the 2014 ACA Medicaid expansions on health insurance coverage and access to care for low-income women of reproductive age. We further investigate whether these effects vary based on women's parental status or prior state Medicaid eligibility policies expanding health insurance coverage for low-income women. This analysis adds to our understanding of the coverage and access gaps created by long-standing categorical Medicaid eligibility criteria for women and the extent to which the ACA has improved access for a population with much to gain from expanded coverage.
Our analysis is informed by the Andersen model of access to health care, which asserts that the use of health services is influenced by individuals’ predisposing characteristics, enabling characteristics, and need (
Andersen and Aday, 1978
). We hypothesize that the 2014 ACA Medicaid expansions led to increased health insurance coverage for women of reproductive age and that this, in turn, led to improved access to care. Owing to the categorical nature of Medicaid eligibility before the 2014 expansion, we anticipate that women without dependent children will experience greater effects than those with dependent children. Finally, we expect that women residing in states that had less generous Medicaid eligibility policies before the ACA will experience greater effects of Medicaid expansion than women of reproductive age residing in states with more generous pre-ACA policies.Methods
Data and Study Sample
This study uses secondary data from the Behavioral Risk Factor Surveillance System (BRFSS) to estimate the effects of the 2014 ACA Medicaid expansions on women of reproductive age. We analyze data from 2012 to 2015, which includes a 2-year (2012 and 2013) preexpansion period and a 2-year (2014 and 2015) postexpansion period. The BRFSS is a system of telephone surveys of more than 400,000 adult (≥18 years of age) U.S. residents with questions on health-related risk behaviors, chronic health conditions, and the use of preventive services. Samples are chosen to be state representative; a common set of questions are used across states, with flexibility for states to supplement their survey. BRFSS data, including survey weights, are publicly available through the Centers for Disease Control and Prevention. Data from the Area Health Resource File on state unemployment and the Health Resources and Services Administration on federally qualified health centers are merged at the state level.
Our analytic study sample includes women of reproductive age (19–44 years) with household incomes of less than 100% of the FPL in 14 states that expanded Medicaid on January 1, 2014, and 16 states without an ACA Medicaid expansion. We omit 18-year-olds from our analysis owing to their eligibility for coverage through the Children's Health Insurance Program, independent of the ACA. Because pregnancy among women with incomes of less than 100% of the FPL would qualify them for Medicaid before the ACA, we omit those women reporting they were pregnant at the time of the BRFSS survey from our analytic sample. This exclusion applies to approximately 4.7% of women with dependent children and 3.3% of women without dependent children in the total sample. The resulting analytic samples vary slightly by dependent variable, ranging from 24,955 to 25,816 low-income, nonpregnant women of reproductive age.
Twenty states and the District of Columbia are excluded from this analysis owing to early or late ACA Medicaid expansions or the presence of pre-ACA Medicaid waiver covering childless adults up to 100% of the FPL. We limit our sample to expansions that occurred on January 1, 2014, to ensure a balanced analytic sample that includes the 2 years before the expansion (2012–2013) and 2 years after the expansion (2014–2015). Including 2 years of postperiod data, in particular, is important to allow for the identification of lagged effects and effects for BRFSS questions that refer to the entire 12-month period before the interview. A full classification of states is presented in Table 1. Although expansion increased Medicaid eligibility to 138% of the FPL (133% FPL with the addition of a 5% disregard), we limit our sample to those with incomes below 100% of the FPL because women with incomes 100% to 138% of the FPL in nonexpansion states were eligible for health insurance subsidies through the Marketplace beginning in January 2014.
Table 1Classification of States by Medicaid Expansion Decision
| Expansion States | Nonexpansion States | Excluded States |
|---|---|---|
| Arkansas | Alabama | Alaska |
| Colorado | Florida | Arizona |
| Illinois | Georgia | California |
| Iowa | Idaho | Connecticut |
| Kentucky | Kansas | Delaware |
| Maryland | Louisiana | District of Columbia |
| Massachusetts | Mississippi | Hawaii |
| Nevada | Nebraska | Indiana |
| New Mexico | North Carolina | Maine |
| North Dakota | Oklahoma | Michigan |
| Ohio | South Carolina | Minnesota |
| Oregon | South Dakota | Missouri |
| Rhode Island | Texas | Montana |
| West Virginia | Utah | New Hampshire |
| Virginia | New Jersey | |
| Wyoming | New York | |
| Pennsylvania | ||
| Tennessee | ||
| Vermont | ||
| Washington | ||
| Wisconsin |
Note: To better measure the effects of the 2014 Affordable Care Act (ACA) expansions, we omit women in states with prior Medicaid expansions to childless adults through section 1115 waivers or early ACA expansions and those in states that expanded Medicaid after January 1, 2014.
Measures
Our dependent variables are indicators of whether a woman 1) was uninsured, 2) needed to see a doctor, but could not due to cost 3) had no personal doctor, and 4) had a primary care visit in the past year. The first three measures are indicators of a lack of access, whereas having a primary care visit in the past year is a positive indicator of access. All models control for individual variables likely to affect insurance coverage and access: 1) age, 2) race/ethnicity, 3) education, 4) work status, 5) marital status, 6) household size, 7) health status, and 8) urban/rural county. Finally, we include two state-level measures: state unemployment rate accounts for changes in the overall economy and the number of federally qualified health centers per women ages 19 to 44 in poverty accounts for supply of health care services that do not require insurance coverage.
Statistical Analysis
We use 2 years (2012 and 2013) before and 2 years (2014 and 2015) after the implementation of ACA Medicaid expansions to assess the effects of expansion on two groups of women: 1) women with dependent children, and 2) women without dependent children. This classification is based on the number of children in a woman's household.
We apply a quasi-experimental difference-in-differences approach based on the variation in states’ decisions to expand Medicaid using cross-state logistic regression models that include state and year fixed effects. We measure the effect of expansion using a dummy variable indicating the presence of Medicaid expansion in a given state-year. When state fixed effects and year fixed effects are also included in the model, the coefficient of this variable measures the effect of the presence of an expansion on the likelihood of each outcome for women of reproductive age, holding other factors constant. All analyses were conducted using STATA 14 (StataCorp, College Station, TX). Models include BRFSS survey weights and standard errors clustered at the state level; results are reported as marginal effects.
We conduct two sets of secondary analyses of our primary outcome, uninsurance. First, we classify states by their pre-ACA eligibility level to capture differences in the magnitude of expansions. We use the following categories: 1) expansion from a pre-ACA Medicaid eligibility level of less than 50% of the FPL, 2) expansion from a pre-ACA Medicaid eligibility level of 50% or greater of the FPL, and 3) no expansion. Pre-ACA eligibility levels are specific to a woman's parental status; women with dependent children are assigned the pre-ACA parental eligibility level and women without dependent children are assigned the pre-ACA childless adult eligibility level. Second, we estimate separate effects for 1) women residing in states with a Medicaid family planning waiver, and 2) women residing in states without a Medicaid family planning waiver. Arkansas is excluded from this analysis because they implemented a family planning waiver during our study period.
Results
Women's characteristics are generally consistent between the pre- and post-expansion periods (Table 2). Overall, women without dependent children are younger, more educated, more likely to be White, and more likely to be unmarried than women with dependent children. Based on descriptive data alone, rates of uninsurance and reported inability to access care owing to cost were lower in the postexpansion period compared with the pre-expansion period (Table 3). Rates of primary care visits were also higher in the post period, although the reported lack of a personal doctor remained consistent over time. Overall, women without dependent children were less likely to be uninsured but more likely to not have a personal doctor than women with dependent children.
Table 2Characteristics of Low-Income Women of Reproductive Age, by Year and Sample, 2012–2015
| All Women < 100% FPL (%) | Women with Dependent Children < 100% FPL (%) | Women without Dependent Children < 100% FPL (%) | ||||
|---|---|---|---|---|---|---|
| 2012–2013 | 2014–2015 | 2012–2013 | 2014–2015 | 2012–2013 | 2014–2015 | |
| Age | ||||||
| 19–26 | 29.3 | 28.3 | 24.3 | 22.7 | 53.3 | 54.7 |
| 27–34 | 36.9 | 35.5 | 40.7 | 39.7 | 18.2 | 16.0 |
| 35–44 | 33.8 | 36.2 | 35.0 | 37.6 | 28.5 | 29.3 |
| Race/ethnicity | ||||||
| White | 37.2 | 36.8 | 33.9 | 33.5 | 53.3 | 52.5 |
| Black | 24.1 | 23.9 | 24.2 | 23.9 | 23.6 | 23.8 |
| Asian | 2.6 | 3.4 | 2.1 | 2.8 | 5.1 | 6.2 |
| Other | 2.7 | 1.9 | 2.6 | 1.7 | 3.2 | 3.0 |
| Hispanic | 33.4 | 34.0 | 37.2 | 38.1 | 14.8 | 14.5 |
| Educational attainment | ||||||
| Less than high school degree | 32.3 | 31.3 | 34.7 | 34.3 | 20.4 | 17.3 |
| High school degree | 31.9 | 31.1 | 33.1 | 32.1 | 25.8 | 26.4 |
| Some college | 29.8 | 31.0 | 27.4 | 28.3 | 41.6 | 43.5 |
| College degree or more | 6.0 | 6.7 | 4.8 | 5.4 | 12.3 | 12.8 |
| Work Status | ||||||
| Nonworker | 60.2 | 60.0 | 58.0 | 58.2 | 70.7 | 68.1 |
| Worker | 39.8 | 40.0 | 42.0 | 41.8 | 29.3 | 31.9 |
| Marital status | ||||||
| Not married | 69.8 | 68.2 | 66.3 | 64.2 | 86.7 | 86.5 |
| Married | 30.2 | 31.8 | 33.7 | 35.8 | 13.3 | 13.5 |
| Household size | ||||||
| 1 | 14.8 | 15.3 | -- | -- | 86.7 | 86.5 |
| 2 | 16.3 | 16.8 | 17.0 | 17.6 | 13.3 | 13.5 |
| 3 | 23.7 | 21.1 | 28.5 | 25.6 | -- | -- |
| ≥4 | 45.2 | 46.8 | 54.5 | 56.8 | -- | -- |
| Health status | ||||||
| Excellent | 13.5 | 14.4 | 13.1 | 14.2 | 15.5 | 15.4 |
| Very good | 20.1 | 21.5 | 19.3 | 20.8 | 24.0 | 24.6 |
| Good | 40.1 | 39.1 | 42.0 | 40.7 | 31.0 | 32.2 |
| Fair | 20.0 | 20.0 | 19.9 | 20.0 | 20.5 | 19.8 |
| Poor | 6.3 | 5.0 | 5.7 | 4.3 | 8.9 | 8.0 |
| Urban/rural | ||||||
| Rural | 8.0 | 6.2 | 8.3 | 6.3 | 6.5 | 5.7 |
| Urban | 23.3 | 17.7 | 24.6 | 18.4 | 16.9 | 14.2 |
| Missing | 68.7 | 76.1 | 67.1 | 75.3 | 76.6 | 80.1 |
| Unemployment rate | 7.4 | 5.6 | 7.4 | 5.6 | 7.4 | 5.6 |
| Number of FQHC per 100,000 women in poverty | 11.7 | 12.9 | 11.5 | 12.8 | 12.6 | 13.4 |
| Sample size ∗ Sample size and descriptive statistics are reported from the sample for the uninsured model. Samples vary slightly across models depending on missing observations for the dependent variable. All samples include nonpregnant women of reproductive age (19–44 years) with household incomes of less than 100% of the FPL. Years 2012–2013 are before the Affordable Care Act (ACA) Medicaid expansion and years 2014–2015 are after the ACA Medicaid expansion. Data are adjusted by BRFSS survey weights. Samples include women residing in both expansion and nonexpansion states. The definition of the women with dependent children category does not allow for a household size of one. The definition of the women without dependent children category does not allow for a household size of more than two. Owing to the high percentage of missing values for the BRFSS Metropolitan Statistical Area (MSA) variable used to derive urban/rural, this measure includes three categories: urban, rural, and missing. | ||||||
| Unweighted n | 14,169 | 11,617 | 11,765 | 9,693 | 2,404 | 1,924 |
| Weighted n (in millions) | 10.8 | 10.3 | 8.9 | 8.5 | 1.8 | 1.8 |
Abbreviations: FPL, Federal Poverty Level; FQHC, federally qualified health centers.
Note: Data from the Behavioral Risk Factor Surveillance System (BRFSS) 2012–2015.
∗ Sample size and descriptive statistics are reported from the sample for the uninsured model. Samples vary slightly across models depending on missing observations for the dependent variable. All samples include nonpregnant women of reproductive age (19–44 years) with household incomes of less than 100% of the FPL. Years 2012–2013 are before the Affordable Care Act (ACA) Medicaid expansion and years 2014–2015 are after the ACA Medicaid expansion. Data are adjusted by BRFSS survey weights. Samples include women residing in both expansion and nonexpansion states. The definition of the women with dependent children category does not allow for a household size of one. The definition of the women without dependent children category does not allow for a household size of more than two. Owing to the high percentage of missing values for the BRFSS Metropolitan Statistical Area (MSA) variable used to derive urban/rural, this measure includes three categories: urban, rural, and missing.
Table 3Insurance Coverage and Access to Health Care among Low-Income Women of Reproductive Age, by Year and Sample, 2012–2015
| All Women < 100% FPL (%) | Women with Dependent Children < 100% FPL (%) | Women without Dependent Children < 100% FPL (%) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| All | 2012–2013 | 2014–2015 | All | 2012–2013 | 2014–2015 | All | 2012–2013 | 2014–2015 | |
| Uninsured | 43.9 | 48.0 | 39.6 | 45.1 | 48.8 | 41.2 | 38.1 | 44.4 | 31.9 |
| Needed to see doctor but could not due to cost | 40.1 | 43.3 | 36.7 | 40.2 | 43.3 | 37.0 | 39.4 | 43.4 | 35.4 |
| No personal doctor | 40.9 | 41.2 | 40.6 | 40.7 | 40.9 | 40.4 | 42.0 | 42.7 | 41.2 |
| Primary care visit within past year | 60.2 | 58.7 | 61.9 | 60.0 | 58.5 | 61.6 | 61.4 | 59.6 | 63.2 |
| Sample size ∗ Sample size and descriptive statistics are reported from the sample for the uninsured model. Samples vary slightly across models depending on missing observations for the dependent variable. All samples include nonpregnant women of reproductive age (19–44 years) with household incomes of less than 100% of the FPL. Years 2012–2013 are before the Affordable Care Act (ACA) Medicaid expansion and years 2014–2015 are after the ACA Medicaid expansion. Data are adjusted by BRFSS survey weights. Samples include women residing in both expansion and nonexpansion states. | |||||||||
| Unweighted n | 25,786 | 14,169 | 11,617 | 21,458 | 11,765 | 9,693 | 4,328 | 2,404 | 1,924 |
| Weighted n (in millions) | 21.1 | 10.8 | 10.3 | 17.4 | 8.9 | 8.5 | 3.6 | 1.8 | 1.8 |
Abbreviation: FPL, Federal Poverty Level.
Note: Data from the Behavioral Risk Factor Surveillance System (BRFSS) 2012–2015.
∗ Sample size and descriptive statistics are reported from the sample for the uninsured model. Samples vary slightly across models depending on missing observations for the dependent variable. All samples include nonpregnant women of reproductive age (19–44 years) with household incomes of less than 100% of the FPL. Years 2012–2013 are before the Affordable Care Act (ACA) Medicaid expansion and years 2014–2015 are after the ACA Medicaid expansion. Data are adjusted by BRFSS survey weights. Samples include women residing in both expansion and nonexpansion states.
ACA Medicaid expansions decreased the likelihood that low-income women of reproductive age were uninsured by 13.2 percentage points (p < .001; Table 4). This decrease in uninsurance was driven by effects among women without dependent children, who experienced a 27.4 -percentage point reduction in uninsurance (p < .001). Women with dependent children experienced a significant 10.1-percentage point decrease (p < .05). We find a significant reduction of 3.8 percentage points (p < .05) in the likelihood that all women who needed to see a doctor but could not due to cost, but we do not find an effect in either of our subsamples. Women without dependent children were 13.3 percentage points less likely to report not having a personal doctor after ACA Medicaid expansions (p < .01), but we do not find an effect for women with dependent children or all women. We do not find significant effects for the likelihood of having a primary care visit within the past year for any sample. Full results are reported in Appendix A.
Table 4Effect of ACA Medicaid Expansions on Insurance Coverage and Access to Care among Low-Income Women of Reproductive Age 2012–2015
| Measure (%) | All Women < 100% FPL | Women with Dependent Children < 100% FPL | Women without Dependent Children < 100% FPL |
|---|---|---|---|
| Uninsured | −13.2*** | −10.1* | −27.4*** |
| Needed to see doctor but could not owing to cost | −3.8* | −3.4 | −8.6 |
| No personal doctor | −1.2 | 1.5 | −13.3** |
| Primary care visit within past year | −0.9 | −1.7 | 3.4 |
Abbreviation: FPL, Federal Poverty Level.
Note: Data from the Behavioral Risk Factor Surveillance System (BRFSS) 2012–2015.
*p < .05; **p < .01; ***p < .001.
Sample includes nonpregnant women of reproductive age (19–44 years) with household incomes of less than 100% of the FPL. Logistic regression models adjust for age, race/ethnicity, education, work status, marital status, household size, health status, urban/rural county, state unemployment rate, and state counts of federally qualified health centers per 100,000 women of reproductive age in poverty. Owing to the high percentage of missing values for the BRFSS Metropolitan Statistical Area (MSA) variable used to derive urban/rural, this measure includes three categories: urban, rural, and missing. All models include state fixed effects and year fixed effects and use BRFSS survey weights. Standard errors are clustered at the state level. Results are reported as marginal effects. Models include expansion states (AR, CO, IL, IA, KY, MD, MA, NV, NM, ND, OH, OR, RI, and WV), and nonexpansion states (AL, FL, GA, ID, KA, LA, MS, NE, NC, OK, SC, SD, TX, UT, VA, and WY). Sample sizes vary by dependent variable. All uninsured women, n = 25,786; women with dependent children, n = 21,458; women without dependent children, n = 4,328. Needed to see doctor but could not owing to cost: all women, n = 25,816; women with dependent children, n = 21,472; women without dependent children, n = 4,344. No personal doctor: all women, n = 25,797; women with dependent children, n = 21,451; women without dependent children, n = 4,346. Primary care visit within past year: all women, n = 24,955 N; women with dependent children, n = 20,756; women without dependent children, n = 4,199.
As expected, states’ pre-ACA Medicaid eligibility policies affected the impact of ACA Medicaid expansions on all women of reproductive age (Table 5). For all women, expansions in states with prior eligibility levels less than 50% FPL had a larger effect (19.4-percentage point decrease; p < .001) than those in states with prior eligibility levels of 50% or greater of the FPL (14.2-percentage point decrease; p < .01). The effect for women with dependent children in states with low pre-ACA Medicaid eligibility levels (26.6-percentage point decrease; p < .001) was more than twice the effect for those in states with higher pre-ACA eligibility (11.7-percentage point decrease; p < .05). Effects for all women with incomes of less than 100% of the FPL were also greater in states without a family planning waiver (17.6-percentage point decrease; p < .01) compared with those with a waiver (13.2-percentage point decrease; p < .01).
Table 5Effect of ACA Medicaid Expansions on Insurance Coverage among Low-Income Women of Reproductive Age 2012–2015, by Prior State Eligibility Levels and Family Planning Waivers
| Uninsurance (%) | |||
|---|---|---|---|
| All Women < 100% FPL | Women with Dependent Children < 100% FPL | Women without Dependent Children < 100% FPL | |
| Pre-ACA Medicaid eligibility level | |||
| Pre-ACA Medicaid eligibility < 50% FPL | −19.4*** | −26.6*** | −28.2*** |
| Pre-ACA Medicaid eligibility ≥ 50% FPL | −14.2** | −11.7* | -- |
| Medicaid family planning waivers | |||
| Medicaid family planning waiver | −13.2** | −9.0* | −27.1*** |
| No Medicaid family planning waiver | −17.6** | −14.5* | −26.4*** |
Abbreviations: ACA, Affordable Care Act; FPL, Federal Poverty Level.
Note: Data from the Behavioral Risk Factor Surveillance System (BRFSS) 2012–2015.
*p < .05; **p < .01; ***p < .001.
All samples include nonpregnant women of reproductive age (19–44 years) with household incomes of less than 100% of the FPL. Logistic regression models adjust for age, race/ethnicity, education, work status, marital status, household size, health status, urban/rural county, state unemployment, and state counts of federally qualified health centers per 100,000 women of reproductive age in poverty. Owing to the high percentage of missing values for the BRFSS Metropolitan Statistical Area (MSA) variable used to derive urban/rural, this measure includes three categories: urban, rural, and missing. All models include state fixed effects and year fixed effects and use BRFSS survey weights. Standard errors are clustered at the state level. Results are reported as marginal effects. Uninsurance by pre-ACA Medicaid eligibility level: all women, n = 25,786; women with dependent children, n = 21,458; women without dependent children, n = 4,328. Uninsurance in states with a Medicaid family planning waiver: all women, n = 10,791; women with dependent children, n = 9,086; women without dependent children, n = 1,705. Uninsurance in states without a Medicaid family planning waiver: all women, n = 14,003; women with dependent children, n = 11,547; women without dependent children, n = 2,456. Expansion states with prior eligibility levels of less than 50% of the FPL for women with dependent children: AR, OR, and WV. Expansion states with prior eligibility levels of 50% or greater of the FPL for women with dependent children: CO, IA, KY, NV, ND, and OH. Expansion states with prior eligibility levels of less than 50% of the FPL for women without dependent children: AR, IL, IA, KY, MD, MA, NV, NM, ND, OH, OR, RI, and WV. Expansion states with prior eligibility levels of 50% or greater of the FPL for women without dependent children: none. Nonexpansion states: AL, FL, GA, ID, KA, LA, MS, NE, NC, OK, SD, TX, UT, VA, and WY. Expansion states with family planning waivers: IA, MD, NM, OH, and OR. Nonexpansion states with family planning waivers: AL, GA, LA, MS, SC, TX, and VA. Expansion states without family planning waivers: CO, KY, MA, NV, ND, RI, and WV. Nonexpansion states without family planning waivers: FL, ID, KA, NE, NC, OK, SD, UT, and WY.
Decreases in uninsurance among women without dependent children, however, are generally consistent and large in magnitude across all classifications of states in Table 5, ranging from 26.4 to 28.2 percentage points (p < .001). These effects for women without dependent children were greater than any decreases found for all women or women with dependent children.
Sensitivity Analysis
We conducted tests of the parallel trends assumption of a difference-in-differences analysis. First, we plot trends in our study outcomes over time for women of reproductive age in expansion and nonexpansion states (Appendix B). These trends suggest that the means for each of our four outcomes are generally parallel between expansion and nonexpansion states during the 2012 and 2013 pre-expansion period of our study. We also conduct a formal test for the significance of pre-expansion period trends by interacting our expansion indicator variable with the year dummy for 2013 to test for differences from 2012, the pre-expansion period trend, for each of our four outcome variables (Appendix Table B1). We find no significant differences (p < .05) in the pre-expansion period trends for all women or for our subgroups of women across all outcomes.
Discussion
ACA Medicaid expansions decreased uninsurance among women of reproductive age with incomes below 100% FPL by 13.2 percentage points, on average. Much greater effects were found for women who were categorically excluded from Medicaid eligibility before the ACA, particularly women without dependent children, who experienced a 27.4-percentage point decrease in uninsurance. For all women and for women with dependent children, we also find greater effects for those residing in states that had not expanded Medicaid eligibility for parents with incomes above 50% of the FPL or had not implemented Medicaid family planning programs before the ACA. We find two effects on access. Women without dependent children were 13.3 percentage points less likely to report not having a personal doctor after the ACA Medicaid expansion. Finally, all women were 3.8 percentage points less likely to report not seeing a doctor owing to cost, although this effect was not found in either subgroup of women.
We find a greater reduction in uninsurance for low-income women of reproductive age than the recently published work by
Simon et al., 2017
, which also used BRFSS data. This discrepancy is likely due to differences in our classification of states and our restriction of the sample to women of reproductive age. Our larger estimate is based on analysis that includes only expansions that occurred on January 1, 2014. States with early expansions, who are excluded from our analysis, built off of existing Medicaid eligibility expansions for low-income adults; based on our findings for states with more generous pre-ACA eligibility, these expansions likely had smaller impacts on coverage for women of reproductive age than the average effects found in this study (Sommers et al., 2013
, Sommers et al., 2014
). The inclusion of late expansions may further contribute to Simon's smaller findings, because these states would have a smaller postexpansion period and analysis would, therefore, be unable to identify gains in coverage that may have occurred in the second year of implementation.Increased health insurance coverage among women of reproductive age is an important finding for women's health, women's overall well-being, and the health and well-being of their children. Women who gain coverage may experience the effects of gaining health insurance found in earlier studies, such as reduced all-cause mortality, improved self-reported physical and mental health, increased compliance with recommended preventive services, and decreased medical debt (
Finkelstein et al., 2012
, Sommers et al., 2012
). These improvements in women's health and well-being may also translate to improved health outcomes for their children (Atrash et al., 2006
, Johnson et al., 2015
, Lu and Halfon, 2003
, McCormick, 2001
). Finally, increased health insurance coverage may increase women's access to family planning and contraceptive services, helping women and their partners to plan pregnancies, potentially leading to health, social, and economic benefits for women and their families (Bailey, 2013
, Sonfield et al., 2013
).Our findings highlight the impact that state flexibility in Medicaid policy design has had on women's health insurance coverage in prior years and, hence, under the ACA. The larger effects of ACA Medicaid expansions found for women residing in states with low pre-ACA eligibility and without family planning waivers suggests that such women faced greater barriers to pre-ACA insurance coverage than their peers in other states. The large effect found for women without dependent children underscores the extent to which these women were previously excluded from Medicaid eligibility, limiting the ability to provide the life course view of health and appropriate services most tightly linked to improved maternal and infant health outcomes (
Adams and Johnston, 2016
). The differences in effects for all women with incomes of less than 100% of the FPL and women with dependent children in states with and without a family planning waiver points to the need to further understand how women interpret the coverage they receive under these waiver programs. If women enrolled in Medicaid family planning waiver programs, who are eligible for only certain preventive and reproductive health benefits, consider themselves to be insured, our results may understate the effect of ACA Medicaid expansions on increasing comprehensive health insurance coverage.Our finding of mixed effects for women's access to care is generally consistent with the earlier work of
Simon et al., 2017
, which found only marginally significant improvements in access for women and no significant changes in use of preventive care or changes in health behaviors. Second-order effects—outcomes that are expected to change owing to changes in health insurance coverage—may not be measurable owing to the limited postexpansion period observed in this and earlier work. Because the current analysis is limited by data availability to the 2 years after the expansion, we are unable to assess whether expansions had effects on access to care later in the implementation period.We do, however, find that ACA Medicaid expansions increased the likelihood that women without dependent children had a personal doctor. Although we find limited effects for experiencing unmet need owing to cost and no effects for having a primary care visit in the past year, free clinics or other sources of care for uninsured women may meet these needs. Women accessing care through safety net systems, however, may be less likely to consider themselves to have a personal doctor, even if they have a visit with a health care provider. Having a usual source of care is clinically important, because it is associated with improved access to care, including increased receipt of preventive care and screening services and reduced access barriers and reported unmet need (
Blewett et al., 2008
, DeVoe et al., 2003
, Shi et al., 2013
).Future research should continue to investigate the effects of ACA Medicaid expansions on access to care as additional years of postexpansion period data become available. Researchers should consider the extent to which eligibility expansions may have resulted in overcrowded provider networks, thereby mitigating any expected improvements in access owing to increased coverage (
Miller and Wherry, 2017
), although an initial audit study suggests that primary care access for Medicaid patients increased between 2012 and 2016 (Polsky et al., 2017
). Here too, qualitative research could shed light on additional factors influencing women's access to care after Medicaid expansions. For example, newly insured women may face challenges interacting with the health care system, particularly for planned and preventive health services. Conversely, before Medicaid, uninsured women may have found sources of care that did not require health insurance, such as federally qualified health centers or more women's-health centered sources such as Title X clinics, Planned Parenthood clinics, or through grant-funded maternity care (Title V). Finally, additional research is needed to assess whether the ACA Medicaid expansions mitigated disparities in coverage and access for women of reproductive age and to identify populations who continue to face barriers. As more women of reproductive age gain health insurance coverage, additional outreach and education may be needed to ensure that increased coverage translates into improved access for all groups.Limitations
The primary limitation of this study is the short, 2-year postexpansion period examined. Owing to the limitations of BRFSS data availability, we are unable to capture effects of Medicaid expansion that occurred more than 2 years after implementation. A second limitation is the inability to measure type of health insurance using BRFSS data. This prevents us from analyzing whether the decrease in uninsurance is due to an increase in Medicaid coverage. Further, the BRFSS does not ask questions about participation in Medicaid family planning programs. Therefore, we are unable to capture the extent to which women residing in states with such programs are affected by them. Finally, our interpretation of the reduction in uninsurance must be considered in light of the possibility that women report noncomprehensive coverage, such as that provided under a family planning waiver, as insurance coverage, particularly during the pre-ACA period.
Implications for Practice and/or Policy
As the wide range of policies targeting women's health—including Title X and Planned Parenthood funding, the required coverage of contraception and other women's preventive services without cost sharing under the ACA, and broad changes to the structure and financing of the Medicaid program—are debated at both the federal and state levels, our findings highlight the role of ACA Medicaid expansion as a policy tool to increase health insurance coverage for women of reproductive age. The coverage gains identified in this study remain at risk if Congress repeals or reduces funding for ACA Medicaid expansions or if states roll back their current expansions.
Our findings also highlight the opportunity that remains for states that have yet to expand Medicaid eligibility under the ACA. The majority of these 19 states have eligibility levels of less than 50% of the FPL for parents and no coverage for childless adults (
Kaiser Family Foundation, 2017
), the state policy characteristics we find to be associated with the largest gains in coverage for reproductive-age women owing to ACA Medicaid expansion. In addition to low eligibility levels, more than one-half of these states are located in the southeastern United States, where income is generally lower and maternal and infant outcomes are worse. Although many of these states operate family planning waivers, the women in these states stand to benefit greatly from future Medicaid expansion, providing them with an expanded range of health insurance benefits.Conclusions
ACA Medicaid expansions successfully increased health insurance coverage among low-income women of reproductive age, and had a particularly large effect for women without dependent children. However, we found limited evidence of a change in access to care in the 2-year period immediately after implementation. The large magnitude of the insurance effect for women without dependent children suggests that these women may have faced greater barriers than those with dependent children to health insurance pre-ACA. Such barriers are consistent with the previous categorical eligibility structure of Medicaid, which omitted childless adults, and points to the importance of states’ expansion choices. The large magnitude of the effects found for states expanding from less than 50% of the FPL underscores the opportunity that remains for states yet to expand coverage, particularly those that do not currently operate Medicaid family planning programs. Increasing access for all women of reproductive age is critical to ensuring the health and well-being of women, mothers, and their children in the United States.
Supplementary Data
- Appendices A and B
References
- Insuring women in the United States before, during, and after pregnancies.American Journal of Public Health. 2016; 106: 585
- Access to medical care in the US: Realized and potential.Medical Care. 1978; : 533-546
- The effects of Medicaid expansion under the ACA: Findings from a literature review.The Kaiser Family Foundation, Washington, DC2016 (Available:) (Accessed: March 25, 2017)
- Expanding Medicaid to low-income childless adults under health reform: Key lessons from state experiences.Kaiser Family Foundation, Washington, DC2010
- Preconception care for improving perinatal outcomes: The time to act.Maternal and Child Health Journal. 2006; 10: 3-11
- Fifty years of family planning: New evidence on the long-run effects of increasing access to contraception.National Bureau of Economic Research, 2013 (Available:) (Accessed: January 13, 2017)
- Promising practices in preconception care for women at risk for poor health and pregnancy outcomes.Maternal and Child Health Journal. 2006; 10: S21-28
- When a usual source of care and usual provider matter: Adult prevention and screening services.Journal of General Internal Medicine. 2008; 23: 1354
- Early impacts of the Affordable Care Act on health insurance coverage in Medicaid expansion and non-expansion states.Journal of Policy Analysis and Management. 2017; 36: 178-210
- Saving Babies: The efficacy and cost of recent changes in the Medicaid eligibility of pregnant women.Journal of Political Economy. 1996; 104: 1263-1296
- Receipt of preventive care among adults: Insurance status and usual source of care.American Journal of Public Health. 2003; 93: 786-791
- Changes in prenatal care timing and low birth weight by race and socioeconomic status: Implications for the Medicaid expansions for pregnant women.Health Services Research. 2001; 36: 373-398
- The Oregon Health Insurance Experiment: Evidence from the first year.Quarterly Journal of Economics. 2012; 127: 1057-1106
- The impact of the Medicaid expansions for pregnant women: A synthesis of the evidence.Medical Care Research and Review. 2001; 58: 3-30
- Preconception care: Risk reduction and health promotion in preparation for pregnancy.Jama. 1990; 264: 1147-1149
- Improving Medicaid: Three decades of change to better serve women of childbearing age.Clinical Obstetrics & Gynecology. 2015; 58: 336-354
- Recommendations to improve preconception health and health care – United States: A Report of the CDC/ATSDR Preconception Care Work Group and the Select Panel on Preconception Care (pp. 1–23.Centers for Disease Control and Prevention, Atlanta, GA2006 (Available:) (Accessed: August 1, 2016)
- Medicaid and CHIP income eligibility limits for pregnant women, 2003-2016.2016 (Available:) (Accessed: September 6, 2016)
- Current status of state Medicaid expansion decisions.2017 (Available:) (Accessed: May 17, 2015)
- Racial and ethnic disparities in birth outcomes: A life-course perspective.Maternal and Child Health Journal. 2003; 7: 13-30
- Preconception care between pregnancies: The content of internatal care.Maternal and Child Health Journal. 2006; 10: 107-122
- March 2014 Report to the Congress on Medicaid and CHIP.2014 (Available:) (Accessed: August 9, 2016)
- Prenatal care–Necessary but not sufficient.Health Services Research. 2001; 36: 399-403
- Affordable Care Act Medicaid expansion increased coverage, improved affordability of care, and reduced psychological distress for low-income parents.Health Affairs. 2017; 36: 808-818
- Health and access to care during the first 2 years of the ACA Medicaid expansions.New England Journal of Medicine. 2017; 376: 947-956
- Changes in primary care access between 2012 and 2016 for new patients with Medicaid and private coverage.JAMA Internal Medicine. 2017; 177: 588-590
- A foundation for health reform: Findings of an annual 50-state survey of eligibility rules, enrollment and renewal procedures and cost-sharing practices in Medicaid and CHIP for children and parents during 2009 (Kaiser Commission on Medicaid and the Uninsured No. 8028).Kaiser Family Foundation, Washington, DC2009 (Available:) (Accessed: May 7, 2017)
- Medicaid expansions and cervical cancer screening for low-income women.Health Services Research. 2017; (Available:) (Accessed: September 12, 2017)
- Making the most of Medicaid.Women’s Health Issues. 2008; 18: S41-S46
- Making the most of Medicaid: Promoting the health of women and infants with preconception care.Women’s Health Issues. 2008; 18: S41-S46
- Type of usual source of care and access to care.Journal of Ambulatory Care Management. 2013; 36: 209-221
- The impact of health insurance on preventive care and health behaviors: Evidence from the first two years of the ACA Medicaid expansions.Journal of Policy Analysis and Management. 2017; 36: 390-417
- Lessons from early Medicaid expansions under health reform: Interviews with Medicaid officials.Medicare & Medicaid Research Review. 2013; 3
- Mortality and access to care among adults after state Medicaid expansions.New England Journal of Medicine. 2012; 367: 1025-1034
- Changes in self-reported insurance coverage, access to care, and health under the Affordable Care Act.JAMA. 2015; 314: 366
- New evidence on the Affordable Care Act: Coverage impacts of early Medicaid expansions.Health Affairs. 2014; 33: 78-87
- Estimating the impact of expanding Medicaid eligibility for family planning services: 2011 update.Guttmacher Institute, New York2011 (Available:) (Accessed: January 12, 2017)
- The social and economic benefits of women’s ability to determine whether and when to have children.Guttmacher Institute, New York2013 (Available:) (Accessed: July 11, 2016)
- Early coverage, access, utilization, and health effects associated with the Affordable Care Act Medicaid expansions: A quasi-experimental study.Annals of Internal Medicine. 2016; 164: 795-803
Biography
Emily M. Johnston, PhD, is a Research Associate in the Health Policy Center at the Urban Institute. Her research focuses on access to health care, particularly the effects of state and federal policies on the health and well-being of vulnerable populations.
Andrea E. Strahan, MPP, is a PhD student in the department of Health Policy and Management in the Rollins School of Public Health, Emory University. Her research interests focus on policies that affect access and care delivery for women and children.
Peter Joski, MSPH, is a Senior Associate in the department of Health Policy and Management in the Rollins School of Public Health, Emory University, with a focus on data management and statistical analysis using SAS/Stata.
Anne L. Dunlop, MD, MPH, Associate Professor, Emory University Nell Hodgson Woodruff School of Nursing, is a family and preventive medicine physician and maternal-child health researcher who conducts health services and epidemiologic research in women's preconception and interconception health care.
E. Kathleen Adams, PhD, is Professor in the department of Health Policy and Management in the Rollins School of Public Health, Emory University. Her research is focused on the Medicaid program and other policies affecting access for low-income and vulnerable populations.
Article info
Publication history
Published online: December 20, 2017
Accepted:
November 15,
2017
Received in revised form:
November 14,
2017
Received:
May 18,
2017
Footnotes
Funding Statement: Funded in part by the Ohio Department of Medicaid and through the Medicaid Technical Assistance and Policy Program via federal financial participation funds. The views stated are those of the researchers only and are not attributed to the study sponsors, the Ohio Department of Medicaid or to the Federal Medicaid Program. No financial disclosures were reported by the authors of this paper.
This work was completed while Emily M. Johnston was a PhD candidate at Emory University.
Identification
Copyright
© 2017 Jacobs Institute of Women's Health. Published by Elsevier Inc.
