Enrollment in California’s Medicaid Program After the Affordable Care Act Expansion

Abstract

Objectives. To determine enrollment rates and predictors of enrollment for newly eligible low-income adults in California following the Affordable Care Act’s (ACA’s) Medicaid expansion.

Methods. We used data from the 2014 to 2015 California Health Interview Survey to examine post-ACA Medicaid enrollment rates and multivariable logistic regression to assess the association of demographic factors, income, and health with enrollment.

Results. We found a 78.5% enrollment rate for the newly eligible Medicaid population, translating to 3.8 million adults enrolled and 1.1 million adults who were eligible but did not enroll. Significant predictors of enrollment were participating in a public welfare program (odds ratio [OR] = 6.59; 95% confidence interval [CI] = 3.09, 14.04), having heart disease (OR = 4.03; 95% CI = 1.34, 12.15), being in the top quartile of income (OR = 3.59; 95% CI = 1.64, 7.85), enrolling in 2015 (OR = 3.28; 95% CI = 1.94, 5.56), being unemployed (OR = 2.10; 95% CI = 1.15, 3.82), and being female (OR = 1.71; 95% CI = 1.03, 2.85). We did not find significant disparities across race/ethnicity, education level, or geography.

Conclusions. Repeal of Medicaid expansion would have a substantial effect on health insurance coverage among California’s low-income adults, many of whom report chronic health conditions and no alternative sources of affordable coverage. Future research should examine the mechanisms explaining the higher enrollment rates among California’s Medicaid expansion population.

The Affordable Care Act¹ (ACA) provided states with the option to expand Medicaid coverage to nearly all nonelderly adults with income at or below 138% of the federal poverty level. Historically, most Medicaid programs have restricted eligibility to specific categories, such as children, parents of dependent children, disabled persons, and some elderly populations. As of 2016, 31 states and the District of Columbia expanded coverage to more than 11 million newly eligible adults.²

Not all eligible individuals enroll in Medicaid. Prior to the ACA, estimates of Medicaid enrollment rates ranged from 52% to 81%.^3–5 The extent to which these findings apply to newly eligible populations is not known, because the newly eligible population represents adults traditionally excluded, such as childless adults and parents with incomes higher than previous eligibility thresholds. Enrollment rates are central in understanding the effect of the ACA’s Medicaid expansion and characterizing Medicaid-eligible low-income populations that remain without coverage.

We used 2014 to 2015 California Health Interview Survey data to estimate the enrollment rate among the newly eligible population and predictors of enrollment. California engaged in early expansion in some counties via a Medicaid waiver beginning in July 2011 and then fully expanded under the ACA in January 2014. Enrollment in California’s Medicaid program increased from an average of 7.6 million before the ACA to 12.2 million in November 2016, the largest growth in Medicaid coverage among all states.⁴

METHODS

The California Health Interview Survey is a random-digit-dialed, cross-sectional telephone survey of 40 550 adults.⁵ We restricted the study to those newly eligible after expansion: nonpregnant adults younger than 65 years and those with income at or below 138% of the federal poverty level. To focus on the newly eligible, we excluded individuals who qualified before expansion, such as parents with income less than 100% of the federal poverty level and individuals who reported Social Security Disability Insurance or Supplemental Security Income because of disability.

The main outcome was Medicaid enrollment. The independent variables were age, gender, self-identified race/ethnicity, education, marital status, access to other health insurance plans, participation in public programs (Temporary Assistance for Needy Families [TANF] or food stamps), and the presence or absence of diabetes, heart disease, asthma, and obesity.

We examined associations of each variable with enrollment and used χ² analysis to test for statistical significance. We used multivariable logistic regression with survey jackknife variance estimation and population weights. We used Stata version 13 (StataCorp LP, College Station, TX) for analyses.

RESULTS

From 2014 to 2015, we identified 2591 individuals who were newly eligible for Medicaid. Of these, 2118 enrolled (78.5% enrollment rate after accounting for survey weights). Participation rates were 67.8% in 2014 and 87.6% in 2015. Overall, this represents 3.8 million adults who enrolled and 1.1 million adults who did not. Of these 4.9 million individuals, 92.9% reported having no other source of affordable insurance.

Compared with the newly enrolled group, the eligible but uninsured group was more likely to be young, male, unmarried, and employed (Table 1). Although employment rates were higher in the uninsured group, these individuals were less likely to be in the top quartile of income and less likely to participate in other public welfare programs. The uninsured group had lower rates of clinical diseases. We did not find significant disparity in Medicaid enrollment rates by race/ethnicity, education level, geographic region, English proficiency, or access to the Internet.

TABLE 1—

Demographic, Health, and Income Characteristics by Medicaid Enrollment: California Health Interview Survey, 2014–2015

	Medicaid Enrolled, No. or % (95% CI)	Eligible but Not Enrolled, No. or % (95% CI)	P
No. of individuals
Unweighted sample	2 118	473
Weighted population	3 827 791	1 050 240
Age, y
18–34	43.14 (39.46, 46.81)	58.20 (45.41, 71.00)	.031
35–65	56.88 (53.19, 60.54)	41.80 (29.00, 54.59)	.031
Gender
Male	40.50 (36.24, 44.76)	53.18 (43.74, 62.64)	.014
Female	59.51 (55.24, 63.76)	46.82 (37.36, 56.26)	.014
Race/ethnicity
White	21.80 (18.79, 24.82)	22.02 (14.29, 29.74)	.96
African American	7.02 (4.49, 9.55)	4.64 (1.80, 7.45)	.24
Asian American	12.18 (8.19, 16.18)	11.80 (6.59, 17.04)	.91
Hispanic	56.15 (51.84, 60.45)	58.57 (49.79, 67.35)	.66
American Indian/Alaska Native	0.52 (0.15, 0.89)	0.44 (–0.09, 0.96)	.8
Marital status: married	27.63 (21.35, 33.90)	15.42 (7.24, 23.64)	.023
Attained high school diploma	70.28 (66.25, 74.30)	76.73 (67.27, 86.25)	.21
English proficiency	71.68 (67.41, 75.95)	77.80 (69.43, 86.23)	.25
Internet access	76.32 (70.71, 81.91)	82.54 (75.54, 89.52)	.19
Urban	89.69 (87.37, 92.00)	92.66 (89.74, 95.64)	.14
Employed	45.56 (40.62, 50.49)	70.65 (61.21, 80.13)	< .001
Self-perceived health
Excellent	31.13 (26.90, 35.37)	43.80 (32.80, 54.84)	.035
Good	32.88 (28.69, 37.06)	29.96 (20.55, 39.36)	.57
Poor	35.99 (32.38, 39.60)	26.24 (16.86, 35.60)	.08
Clinical health
Obese (BMI > 30 kg/m2)	33.35 (28.16, 38.54)	22.71 (14.28, 31.17)	.049
Diabetes	14.18 (11.17, 17.18)	6.32 (0.00, 13.09)	.11
Heart disease	4.64 (2.64, 6.64)	1.58 (0.19, 2.98)	.022
Asthma	15.20 (11.77, 18.63)	11.61 (6.14, 17.06)	.32
≥ 1 comorbidity reported	29.00 (24.81, 33.18)	18.81 (10.77, 26.88)	.039
Use of public programs	33.61 (29.20, 38.01)	6.27 (2.20, 10.36)	< .001
Income
1st quartile	22.15 (18.64, 25.66)	18.95 (10.98, 26.91)	.49
2nd quartile	25.36 (20.97, 29.75)	32.70 (23.57, 41.88)	.16
3rd quartile	22.17 (17.99, 26.36)	36.47 (27.43, 44.90)	.002
4th quartile	30.32 (25.20, 35.43)	12.17 (5.14, 19.20)	.001

Note. BMI = body mass index; CI = confidence interval.

In multivariable analyses, use of public welfare programs was the strongest predictor of Medicaid enrollment (odds ratio [OR] = 6.59; 95% confidence interval [CI] = 3.09, 14.04). Other significant predictors were having heart disease (OR = 4.03; 95% CI = 1.34, 12.15), being in the top quartile of income (OR = 3.59; 95% CI = 1.64, 7.85), enrolling in 2015 (OR = 3.28; 95% CI = 1.94, 5.56), being unemployed (OR = 2.10; 95% CI = 1.15, 3.82), and being female (OR = 1.71; 95% CI = 1.03, 2.85).

DISCUSSION

Nearly 80% of newly eligible adults participated in California’s Medicaid expansion during the first 2 years of ACA implementation. This rate is substantially higher than the pre-ACA enrollment rates of 51% and 68.5% reported by Davidoff et al.⁶ and Kenney et al.,⁷ respectively, likely reflecting the substantially different characteristics and policy context of the postexpansion population as compared with categorically eligible persons. Our calculated participation rates are also higher than national forecasts of take-up among the expansion population.^8–10

To facilitate enrollment, California’s Medicaid program implemented online, mail, and in-person applications. California also provided temporary coverage to those in the application process, such as those using hospital services, and focused on at-risk populations, automatically qualifying those in the Supplemental Nutrition Assistance Program (SNAP). These measures simplified and shortened the application process,¹¹ and may have contributed to the higher Medicaid enrollment rates among this newly eligible population. Furthermore, the streamlined programs toward vulnerable populations may explain the association of unemployment rate and participation in other public welfare programs with Medicaid enrollment. Other studies have reported that higher income and unemployment were associated with Medicaid take-up.^6,7 Our finding that participation in TANF and SNAP is strongly associated with Medicaid enrollment also aligns with previous research⁷ and has important implications, given policy efforts to reduce spending on these public assistance programs. Lowering the funding to these programs and repealing Medicaid expansion may affect the same disadvantaged population, simultaneously reducing income and access to health care.

Our study suggests that a substantial number of low-income Californians are at risk for losing health insurance coverage if the ACA’s Medicaid expansion is repealed, because 93% of the Medicaid-eligible adults in our sample reported having no other source of affordable coverage. Reversal of the expansion would erode access for a low-income population with more chronic disease and worse self-reported health than in the general population.¹² Because enrollees often churn in and out of Medicaid coverage because of fluctuating incomes, economic downturns, changes to individual health, and policy changes, the number of individuals potentially affected by repeal of Medicaid expansion is likely to be much higher than the estimated 3.8 million enrolled during the time of the survey.

Our study had limitations. California Health Interview Survey’s cross-sectional design prevents making causal inferences about respondents’ characteristics and Medicaid enrollment. Because we could not distinguish between preexisting conditions and conditions diagnosed after enrollment, we selected chronic conditions that were unlikely to change within 1 year of insurance. The survey did not assess whether an individual ever qualified for Medicaid; therefore, our definition of “newly eligible” may include some individuals who qualified before the ACA under eligibility criteria for children, parents, or the disabled. Finally, our findings may not apply to Medicaid programs in other states.

In summary, in the first 2 years of the ACA-initiated Medicaid expansion in California, enrollment rates were 67.8% in 2014 and 87.6% in 2015. Some of the predictors of Medicaid enrollment for post-ACA populations, such as enrollment in other public assistance programs, were similar to those described for pre-ACA enrollment, whereas other factors, such as race/ethnicity, English proficiency, and education, were not associated. Future research should examine the mechanisms explaining the higher enrollment rates among California’s Medicaid expansion population and their applicability to other state Medicaid programs.

PUBLIC HEALTH IMPLICATIONS

Medicaid expansion in California successfully affected a large and growing proportion of the nonelderly, low-income adult population. Changing Medicaid policy to pre-ACA eligibility criteria would likely lead to coverage losses for a substantial number of low-income Californians, many of whom report chronic health conditions and no alternative sources of available coverage.