Racial/Ethnic variation in health insurance coverage and early midlife depressive symptoms in the U.S.

Abstract

Background

Health insurance access is a critical determinant of mental health, yet racial/ethnic disparities in coverage persist. Examining the relationship between health insurance access and depressive symptoms at early midlife can inform interventions addressing mental health inequities.

Methods

To examine the association between health insurance coverage type (private, public, or uninsured) and depressive symptoms (CES-D-5 scale) at early midlife and variation across racial and ethnic groups. We used data from Waves I, IV, and V of the National Longitudinal Study of Adolescent to Adult Health (Add Health), a nationally representative cohort study conducted from 1994 to 2018 in the U.S. The analytic sample included 7,302 respondents who completed Waves I, IV, and V, had valid sample weights, and were not missing key independent or dependent variables. We used ordinary least squares regression to evaluate the association between health insurance type and depressive symptoms.

Results

The analytic sample was 49% women, 4% Asian, 14% Black, 9% Hispanic, and 73% White, and were 38 years old (n = 7,302). Among respondents, 75% had private insurance, 16% had public insurance, and 9% were uninsured. Relative to those with private insurance, adults with public insurance (B = 0.82; 95% CI = 0.53, 1.10; P = < 0.001) and those uninsured (B = 0.95; 95% CI = 0.60, 1.29; P = < 0.001) had significantly greater depressive symptoms. Racial and ethnic differences emerged: Hispanic and White adults with public insurance had greater depressive symptoms, while uninsured White adults reported significantly greater depressive symptoms.

Conclusions

Disparities in health insurance access were associated with differences in depressive symptoms at early midlife, with variation across race and ethnicity. Addressing disparities in access to private health insurance may help improve mental health outcomes.

Background

Early midlife (ages 33–44) is a critical period marked by significant life transitions, including career advancement, family obligations as a caregiver to children and, potentially, older relatives [1]. During this period, individuals often navigate increasing career demands while simultaneously managing increased family responsibilities, including caregiving for both children and aging relatives [1]. These co-occurring stressors may have profound influence for both current and future health and emotional well-being. Despite the importance of this life stage, scholars have frequently understudied this demographic transition in relation to other periods, such as childhood, adolescence, and late life.

From a life course perspective, early midlife represents a pivotal inflection point at which the cumulative impact of early life experience coincides with increasing caregiving and financial responsibilities [2, 3]. These past and present stressors interact to heighten vulnerability to psychological distress and chronic disease [4]. Specifically, early midlife illustrates a key life course stage in which the impact of early life adversity emerges more concretely and where future trajectories of health and wellbeing begin to solidify [5, 6]. Previous longitudinal studies demonstrate that adverse childhood experiences (ACEs) and economic disadvantage were associated with depression, poorer cardiometabolic health (e.g. diabetes and obesity) and chronic inflammation in midlife [7, 8]. Moreover, this life stage and earlier life experiences set the foundation for physical, mental, and cognitive health in later life. For example, findings from the Midlife in the United States (MIDUS) study demonstrated that those with higher psychological wellbeing in early midlife have a lower risk of developing chronic inflammation and related physical health conditions and reporting better overall health in later life [9]. Collectively, this evidence highlights early midlife as a sensitive period when cumulative stressors may have lasting impacts on health and wellbeing across the lifespan.

Given these competing demands and the accumulation of life course stressors, early midlife may be a period of elevated risk for mental health challenges, specifically depression. Access to healthcare plays a crucial role in both the prevention and management of depression and its long-term consequences, including reduced risks of suicide, substance abuse, disability, and co-morbidities [10–12]. For many in the United States, health insurance facilitates access to health care that may otherwise be financially prohibitive. However, available insurance options and benefits are largely limited by federal and state health policy. Within these constraints, individual and family coverage is patterned by sociocontextual factors including income, marital status, employment status, and number of dependents—all of which vary across racial and ethnic groups.

Nationally, the majority of individuals have private health insurance (77%), followed by public health insurance (14%), and 9% are uninsured [13]. However, race- and ethnicity-specific estimates provide a more nuanced description of health insurance coverage in the U.S [14, 15]. Compared to White (6.5%) and Asian (5.8%) individuals aged 0 to 64, higher uninsured rates are observed among Black (9.7%), Hispanic (17.9%), and American Indian and Alaska Native (AIAN) (18.7%) populations. Differences in type of insurance coverage among racial ethnic groups are also notable. Private insurance coverage is more common among White (79.5%) and Asian (80.7%) populations compared to Black (62.9%), Hispanic (56.8%), and AIAN (47.5%) populations. These differences are also notable at early mid-life. For example, a greater share of Asian and White early midlife adults has access to private health insurance (82% and 78%, respectively) relative to Black and Hispanic early midlife adults (67% and 75%, respectively) [16]. Conversely, a greater share of Black and Hispanic early midlife adults have public health insurance or are uninsured compared to Asian and White early midlife adults [16].

These patterns in coverage highlight the importance of considering how insurance type influences access to and quality of mental health care across race/ethnicity. Private, public, and uninsured individuals have increased variability in the mental health providers they can access, with public and uninsured individuals more likely to face barriers to accessing mental health services, thus impacting the quality and effectiveness of their care [17–21]. Individuals with private health insurance are more likely to have greater access to high-quality mental health providers, better continuity of care, and better mental health treatment relative to publicly-insured or uninsured individuals [22–24]. While private insurance has been associated with greater access to specialty care [25], more expansive provider and treatment choice [26–29], and better prognoses [13, 30, 31] compared to public coverage, some studies have highlighted potential disadvantages of private coverage—particularly with mental health care—including higher cost-related barriers [24, 29], lower treatment rates [21, 22], and more unmet mental health care needs [22]. These mixed findings may reflect that the impact of insurance coverage type varies by health care needs, socioeconomic status, and race/ethnicity due to the complex dynamics of structural (i.e., transportation, rurality) and sociocontextual factors (i.e., marital status, immigrant status, wealth and income, health literacy) that serve as barriers to accessing mental health care [32].

These findings also underscore the importance of examining associations between insurance coverage and health-related outcomes across population subgroups. For example, studies among populations with special health care needs found heterogeneity in associations between insurance coverage type and health care access by income [33], race/ethnicity [34, 35], and rurality [34]. A study by Dasgupta et al. found differences in unmet HIV care needs among the privately insured by income—individuals with lower incomes had higher odds of unmet HIV care needs compared to those with higher incomes [33]. Interrante et al. observed intersectional heterogeneity in associations between insurance coverage and postpartum care, with privately insured White urban residents having the highest receipt of care components and privately insured, racially minoritized rural residents having the lowest receipt, with inclusion of those with Medicaid coverage in the analysis [34]. To date, no studies have investigated similar racial/ethnic heterogeneity in the association between health insurance coverage at early midlife and depressive symptoms [36, 37].

Although depression and depressive symptoms are largely preventable, lack of insurance coverage remains a key barrier to accessing timely mental health care for individuals in need [38]. Understanding how race and ethnicity moderate the association between insurance type and depressive symptoms at this critical life course stage is an important first step in developing targeted, equitable approaches to enhance mental health care access. To address this gap, our study leverages nationally representative longitudinal data from the National Longitudinal Study of Adolescent to Adult Health to understand the relationship between health insurance coverage and depressive symptoms at early midlife. We aim to assess how these associations may differ by race and ethnicity to identify patterns of focus for future public health campaigns and interventions focused on bridging the gap between healthcare access and mental well-being.

Methods

Data

For this study, we used Waves I, IV, and V of Add Health and Wave I of the Add Health Parent Questionnaire. Using complete-case analysis [39], we restricted the analytic sample to respondents who had completed Waves I, IV, and V, and did not have missing data on the key independent variables, dependent variables, and did not have missing Wave V sample weights. We also limited our sample to the four largest racial/ethnic groups represented in Add Health: White, Black, Hispanic and Asians. This led to an analytic sample of 7,302 respondents (refer to Fig. 1 for a detailed flow chart). Household income from Wave I of the Add Health Parent Questionnaire had the largest missing data (approximately 11% of the analytic sample).

Fig. 1.

Flow Chart for Analytic Sample

Measures

Key independent variable: early midlife health insurance type

Key Independent Variable: Early Midlife Health Insurance Type. The key independent variable was the respondent’s health insurance type at early midlife. At Wave V of Add Health, respondents were asked “Which of the following best describes your current health insurance situation?” Responses were then categorized into the following groups: private health insurance, including Health Insurance Marketplace or Exchange (1), public health insurance, including from Medicaid, Medicare, TRI-CARE, CHAMPUS, CHAMP-VA, the Department of Veterans Affairs, and the Indian Health Service (2), and no health insurance (3).

Covariates

We included the following covariates, as they have been linked to depressive symptoms.

Gender was whether the respondent identified as a man (1) or woman (2) at Wave I of Add Health.

Race and ethnicity included whether the respondent identified as White (1), Black (2), Hispanic (3), and Asian, including Pacific Islander (4) at Wave V of Add Health.

Family structure at Wave I wasbased on Add Health’s constructed family structure variable ² and included two biological parents (1), two parents (2), a single parent (3), and other (4).

Age was the age of the respondent in adolescence, based on the year the respondent was born subtracted from the year of interview at Wave I of Add Health.

Parental education was the highest level of education that the respondent’s parent had completed at Wave I’s Add Health Parent Questionnaire based on the question, “How far did you go in school?,” and included completed high school or less (1), completed some college (2), completed a college degree (3), and completed more than a college degree (4).

Poor self-rated health was based on the question asked at Wave I of Add Health, “How is your general physical health?” Responses included poor, fair, good, very good, and excellent. We dichotomized responses such that excellent, very good, and good responses were coded as 0 and fair and poor responses were coded as 1.

Depressive symptoms were based on the CES-D-5 scale, [40] and included the following questions, asked of how the respondent felt in the past 7 days at Wave I of Add Health: felt that they could not shake off the blues, even with help from their family and their friends; felt depressed; felt sad; felt life was not worth living; and was happy (reverse-coded). The Cronbach’s alpha for this measure was 0.77. Responses included never or rarely (0), sometimes (1), a lot of the time (2), and most of the time or all of the time (3). Responses were summed and scores ranged from 0-15.

Poor parental self-rated health came from the following question asked at Wave I of the Add Health Parent Questionnaire: “How is your general physical health?” Responses included excellent (1), very good (2), good (3), fair (4), and poor (5). Responses were then dichotomized into excellent, very good, and good health (0) and fair and poor health (1).

Household income came from the following question asked at Wave I of the Add Health Parent Questionnaire: “About how much total income, before taxes did your family receive in 1994? Include your own income, the income of everyone else in your household, and income from welfare benefits, dividends, and all other sources.” Responses ranged from $0 to $999 thousand.

Parental employment status came from the following question asked at Wave I of the Add Health Parent Questionnaire: “Do you work outside the home?” Responses included no (0) and yes (1).

Health insurance type was from Wave I of the Add Health Parent Questionnaire. We included respondents’ health insurance coverage in adolescence from the question asked of their child, “What kind of health insurance does {NAME} have?” We categorized responses into private (1), public (2), and uninsured (3).

Parenthood status came from the following question asked at Wave IV of Add Health: “How many live births resulted from this pregnancy/these pregnancies?” Responses ranged from 0 to 11 live births.

Educational attainment was the respondent’s highest level of educational attainment at Wave IV of Add Health, and included less than high school (1), high school (2), some college (3), and college and more (4).

Full-time employment status was derived from the following questions from Wave V of Add Health: “Are you currently working for pay?” and “How many total hours do you spend at [your job/all your jobs?]” Responses included being employed full-time, working at least 35 hours per week (1) or not (0).

Immigrant generation status was whether respondents were born in the U.S. (0) or were foreign-born (1) from Wave V of Add Health.

Key Dependent Variable: Depressive Symptoms at Wave V. Respondent’s depressive symptoms at Wave V was based on the CES-D-5 scale. [40] This included whether the respondent reported experiencing the following in the past 7 days: felt that they could not shake off the blues, even with help from their family and their friends; felt depressed; felt sad; felt life was not worth living; and was happy (reverse-coded). The Cronbach’s alpha for this measure was 0.82. Responses included never or rarely (0), sometimes (1), a lot of the time (2), and most of the time or all of the time (3). Responses were summed and scores ranged from 0-15.

Analytic strategy

Our primary analytic strategy was using ordinary-least squares regressions to examine how different types of early midlife health insurance was associated with depressive symptoms at early midlife. We also conducted analyses separately across race and ethnicity, to assess whether there was variation in associations of health insurance coverage and depressive symptoms among Asian, Black, Hispanic and White respondents. We used Wave V sample weights to make the results nationally representative of the U.S. population.

Results

Table 1 displays summary statistics of the analytic sample and across race and ethnicity. Among the overall sample, approximately 75% of respondents reported having private health insurance at early midlife, followed by 16% having public health insurance, and 9% being uninsured. Across race and ethnicity, Asian adults had the largest proportion enrolled in private health insurance (80%), followed by White adults (77%), Hispanic adults (73%), and Black adults (66%). Black adults had the largest proportion enrolled in public health insurance (21%), followed by Hispanic adults (16%), White adults (15%), and Asian adults (12%). Black adults also had the largest proportion who were uninsured (13%), followed by Hispanic adults (11%), White adults (8%), and Asian adults (8%).

Table 1.

Summary statistics for analytic sample and across race and ethnicity

Variable	Overall	Asian	Black	Hispanic	White	Range
Health insurance (Wave V)
Private	0.75	0.80b	0.66a, c,d	0.73b	0.77b	1 to 3
Public	0.16	0.12b	0.21a, d	0.16	0.15b
Uninsured	0.09	0.08	0.13d	0.11	0.08b
Gender (Wave I)						1 to 2
Men	0.51	0.51	0.46d	0.50	0.51b
Women	0.49	0.49	0.54d	0.50	0.49b
Race and ethnicity (Wave V)*						1 to 4
Asian	0.04	1.00	0.00	0.00	0.00
Black	0.14	0.00	1.00	0.00	0.00
Hispanic	0.09	0.00	0.00	1.00	0.00
White	0.73	0.00	0.00	0.00	1.00
Family structure (Wave I)						1 to 4
Two biological parents	0.58	0.59b	0.34a, c,d	0.59b	0.63b
Two parents	0.17	0.13	0.14d	0.16	0.17b
Single parent	0.21	0.23b	0.42a, c,d	0.22b	0.17b
Other	0.03	0.04b	0.10a, c,d	0.04b	0.02b
Immigrant generation status (Wave V)						1 to 2
Born in U.S.	0.96	0.73b, d	0.98a, c	0.79b, d	0.99a, c
Foreign-born	0.04	0.27b, d	0.02a, c	0.21b, d	0.01a, c
Age (Wave I)	15.77	15.72	15.96d	15.74	15.74b	12 to 21
Number of siblings (Wave I)	2.47	2.76d	2.48c	2.95b, d	2.39a, c	1 to 15
Parental Education (Wave I)						1 to 4
High school and less	0.45	0.30b, c,d	0.49a, c,d	0.64a, b,d	0.43a, b,c
Some college	0.31	0.27	0.31c	0.26b, d	0.31c
College	0.15	0.29b, c,d	0.11a, c,d	0.07a, b,d	0.16a, b,c
College and more	0.10	0.13b, c	0.08a, c,d	0.04a, b,d	0.10b, c
Household income at Wave I	47.39	55.36b, c	31.43a, d	34.25a, d	51.67b, c	0 to 999
Parent worked at Wave I	0.75	0.79c	0.75	0.70a, d	0.75c	0 to 1
Poor parental self-rated health at Wave I	0.12	0.11b, c	0.18a, d	0.23a, d	0.10b, c	0 to 1
Poor self-rated health (Wave I)	0.06	0.05	0.07	0.07	0.06	0 to 1
Depression at Wave I (CES-D-5)	2.35	2.69d	2.57d	2.79d	2.23a, b,c	0 to 15
Educational attainment (Wave IV)						1 to 4
Less than high school	0.23	0.13b, c,d	0.30a, d	0.28a, d	0.21a, b,c
High school	0.44	0.40	0.45	0.49d	0.43c
Some college	0.21	0.30b, c	0.13a, d	0.15a, d	0.23b, c
College and more	0.12	0.17c	0.12c	0.08a, b,d	0.12c
Parenthood (Wave IV)	0.48	0.29b, c,d	0.60a, c,d	0.51a, b	0.46a, b	0 to 1
Health insurance in adolescence (Wave I)
Private	0.79	0.77b, c	0.63a, d	0.59a, d	0.85b, c
Public	0.09	0.09b	0.23a, c,d	0.13a, b	0.06b, c
Uninsured	0.12	0.13c	0.14a, c	0.28a, b,d	0.09b, c
Employed full-time (Wave V)	0.73	0.68	0.70	0.75	0.73	0 to 1
Depressive symptoms (Wave V) (CES-D-5)	2.39	2.13b	2.63a, d	2.45	2.35b	0 to 15
n	7302	434	1351	893	4624

^a Statistically significant differences from White respondents at p < 0.05 level

^b Statistically significant differences from Black respondents at p < 0.05 level

^c Statistically significant diferences from Hispanic respondents at p < 0.05 level

 ^d Statistically significant differences from Asian respondents at p < 0.05 level

* All racial and ethnic compositions are statistically significantly different from each other for each group

Depressive symptoms at early midlife also varied substantially across race and ethnicity. The overall depressive symptom score was 2.39 for the full sample. By race and ethnicity, Black adults reported the highest depressive symptoms (2.63), followed by Hispanic adults (2.45), White adults (2.35), and Asian adults (2.13). Black adults had statistically significantly greater depressive symptoms at early midlife relative to White and Asian adults.

Table 2 shows ordinary-least squares regression results examining the association between health insurance coverage and early midlife depressive symptoms among the analytic sample, with the reference category of respondents with private health insurance at early midlife including all covariates.

Table 2.

OLS regression of health insurance type and depressive symptoms at early midlife

Variable	B	95% CI		P value
Health insurance type at Wave V (reference: private)
Public	0.82	0.53	1.10	< 0.001
Uninsured	0.95	0.60	1.29	< 0.001
Race and ethnicity at Wave V (reference: non-Hispanic White)
Asian	0.04	−0.22	0.30	0.75
Black	−0.05	−0.33	0.22	0.71
Hispanic	−0.25	−0.57	0.08	0.14
Gender at Wave I (reference: men)
Women	0.08	0.37	−0.25	0.09
Family structure at Wave I (reference: two biological parents)
Two parents	−0.01	0.90	−0.24	0.21
Single parent	−0.09	0.45	−0.31	0.14
Other	0.13	0.61	−0.37	0.64
Immigrant generation status at Wave V (reference: born in U.S.)
Foreign-born	−0.12	0.56	−0.55	0.30
Age of child at Wave I	−0.01	0.56	−0.06	0.03
Number of siblings at Wave I	−0.09	−0.15	−0.03	0.004
Parental Education at Wave I (reference: HS and less)
Some college	0.17	−0.02	0.36	0.09
College	0.11	−0.13	0.34	0.37
More than college	0.13	−0.16	0.43	0.38
Household income at Wave I	0.00	0.00	0.00	< 0.001
Parent worked at Wave I	0.12	−0.08	0.31	0.25
Poor parental self-rated health at Wave I	0.10	−0.16	0.36	0.45
Poor self-rated health at Wave I	0.23	−0.13	0.59	0.21
Depression at Wave I (CES-D-5)	0.17	0.13	0.21	< 0.001
Health insurance in adolescence at Wave I (reference: private)
Public	0.24	−0.11	0.58	0.18
Uninsured	0.30	0.02	0.59	0.04
Parenthood (Wave IV)	−0.12	−0.29	0.05	0.18
Educational attainment at Wave IV (reference: < high school)
High school	−0.04	−0.26	0.19	0.75
Some college	−0.34	−0.59	−0.08	0.009
College and more	−0.38	−0.66	−0.09	0.01
Employed full-time (Wave V)	−0.62	−0.83	−0.41	< 0.001
Constant	2.91	2.00	3.83	< 0.001
n	7302

Relative to respondents who held private health insurance at early midlife, respondents who had public health insurance (B = 0.82; 95% CI = 0.53, 1.10; P = < 0.001) and were uninsured (B = 0.95; 95% CI = 0.60, 1.29; P = < 0.001) had significantly greater depressive symptoms at early midlife.

Table 3 shows ordinary-least squares regression results examining the association between health insurance coverage and early midlife depressive symptoms among Asian and Black adults, including all covariates. Among Asian and Black adults, relative to those who held private health insurance at early midlife, there were no statistically significant associations between public health insurance and being uninsured on depressive symptoms at early midlife.

Table 3.

OLS regression of health insurance type and depressive symptoms at early midlife among Asian and black adults

Variable	Asian				Black
	B	95% CI		P value	B	95% CI		P value
Health insurance type at Wave V (reference: private)
Public	−0.45	−1.32	0.43	0.32	0.31	−0.36	0.98	0.37
Uninsured	0.62	−0.34	1.58	0.21	0.36	−0.42	1.14	0.37
Gender at Wave I (reference: men)
Women	0.12	−0.44	0.69	0.67	0.01	−0.45	0.46	0.98
Family structure at Wave I (reference: two biological parents)
Two parents	0.47	−0.40	1.35	0.29	−0.38	−1.01	0.26	0.25
Single parent	0.09	−0.57	0.76	0.78	−0.23	−0.80	0.34	0.42
Other	−0.67	−1.39	0.04	0.06	−0.42	−1.26	0.42	0.33
Immigrant generation status at Wave V (reference: born in U.S.)
Foreign-born	−0.46	−1.10	0.18	0.16	−0.40	−1.23	0.42	0.34
Age of child at Wave I	−0.09	−0.22	0.05	0.22	0.02	−0.12	0.16	0.80
Number of siblings at Wave I	−0.13	−0.23	−0.02	0.02	−0.11	−0.26	0.04	0.15
Parental Education at Wave I (reference: HS and less)
Some college	0.23	−0.47	0.93	0.52	0.59	0.06	1.12	0.03
College	0.27	−0.49	1.02	0.49	0.30	−0.26	0.86	0.30
More than college	0.42	−0.37	1.21	0.30	0.49	−0.12	1.09	0.11
Household income at Wave I	0.00	0.00	0.00	0.42	0.00	−0.01	0.00	0.48
Parent worked at Wave I	0.39	−0.26	1.04	0.24	−0.23	−0.93	0.48	0.53
Poor parental self-rated health at Wave I	−0.27	−1.08	0.55	0.52	0.08	−0.54	0.69	0.81
Poor self-rated health at Wave I	−1.11	−1.97	−0.25	0.01	0.20	−0.57	0.96	0.61
Depression at Wave I (CES-D-5)	0.19	0.09	0.29	< 0.001	0.12	0.04	0.20	0.01
Health insurance in adolescence at Wave I (reference: private)
Public	0.17	−0.85	1.19	0.74	0.03	−0.69	0.74	0.94
Uninsured	0.66	−0.26	1.59	0.16	0.09	−0.73	0.91	0.83
Parenthood (Wave IV)	−0.27	−0.80	0.26	0.32	−0.19	−0.68	0.29	0.44
Educational attainment at Wave IV (reference: less than high school)
High school	−1.01	−1.87	−0.15	0.02	−0.61	−1.22	−0.01	0.05
Some college	−1.17	−2.14	−0.20	0.02	−0.89	−1.62	−0.16	0.02
College and more	−0.93	−2.07	0.21	0.11	−1.13	−1.89	−0.37	0.01
Employed full-time (Wave V)	−0.11	−0.75	0.53	0.74	−0.73	−1.27	−0.19	0.008
Constant	4.24	1.85	6.63	0.01	3.86	1.19	6.53	0.005
n	434				1351

Table 4 shows ordinary-least squares regression results examining the association between health insurance coverage and early midlife depressive symptoms among Hispanic and White adults, including all covariates.

Table 4.

OLS regression of health insurance type and depressive symptoms at early midlife among Hispanic and white adults

Variable	Hispanic				White
	B	95% CI		P value	B	95% CI		P value
Health insurance type at Wave V (reference: private)
Public	0.83	0.04	1.62	0.04	0.93	0.59	1.28	< 0.001
Uninsured	0.25	−0.53	1.04	0.53	1.10	0.66	1.54	< 0.001
Gender at Wave I (reference: men)
Women	−0.15	−0.64	0.35	0.55	−0.09	−0.29	0.11	0.39
Family structure at Wave I (reference: two biological parents)
Two parents	0.66	−0.06	1.37	0.07	−0.05	−0.31	0.21	0.70
Single parent	0.09	−0.52	0.69	0.78	−0.10	−0.39	0.19	0.50
Other	0.83	−0.87	2.53	0.34	0.30	−0.44	1.04	0.43
Immigrant generation status at Wave V (reference: born in U.S.)
Foreign-born	−0.15	−0.72	0.42	0.60	0.48	−0.61	1.57	0.39
Age of child at Wave I	−0.10	−0.22	0.03	0.12	−0.01	−0.06	0.04	0.74
Number of siblings at Wave I	−0.12	−0.25	0.02	0.09	−0.06	−0.14	0.02	0.17
Parental Education at Wave I (reference: HS and less)
Some college	0.25	−0.38	0.89	0.43	0.07	−0.15	0.29	0.55
College	0.18	−0.77	1.14	0.71	0.05	−0.22	0.33	0.70
More than college	−0.78	−1.92	0.36	0.18	0.07	−0.29	0.43	0.69
Household income at Wave I	0.00	0.00	0.00	0.52	0.00	0.00	0.00	0.12
Parent worked at Wave I	0.15	−0.38	0.68	0.58	0.17	−0.06	0.39	0.15
Poor parental self-rated health at Wave I	0.13	−0.41	0.67	0.64	0.11	−0.24	0.46	0.53
Poor self-rated health at Wave I	−0.05	−0.79	0.69	0.89	0.33	−0.13	0.80	0.16
Depression at Wave I (CES-D-5)	0.17	0.07	0.27	0.01	0.18	0.13	0.23	< 0.001
Health insurance in adolescence at Wave I (reference: private)
Public	−0.29	−1.03	0.44	0.43	0.48	0.02	0.95	0.04
Uninsured	0.47	−0.18	1.12	0.16	0.32	−0.03	0.67	0.07
Parenthood (Wave IV)	0.06	−0.45	0.57	0.82	−0.11	−0.31	0.09	0.27
Educational attainment at Wave IV (reference: less than high school)
High school	0.14	−0.41	0.68	0.62	0.11	−0.16	0.38	0.44
Some college	−0.26	−0.92	0.40	0.45	−0.17	−0.47	0.13	0.27
College and more	0.21	−0.83	1.24	0.70	−0.23	−0.57	0.10	0.17
Employed full-time (Wave V)	−1.03	−1.70	−0.36	0.01	−0.61	−0.86	−0.37	< 0.001
Constant	4.33	1.94	6.73	< 0.001	2.00	0.59	3.41	0.006
n	893				4624

Among Hispanic adults, relative to those who held private health insurance at early midlife, Hispanic adults who held public health insurance had significantly greater depressive symptoms at early midlife (B = 0.83; 95% CI = 0.04, 1.62; P = 0.04).

Among White adults, relative to those who held private health insurance at early midlife, White adults with public health insurance (B = 0.93; 95% CI = 0.59, 1.28; P = < 0.001) or were uninsured (B = 1.10; 95% CI = 0.66, 1.54; P = < 0.001) had significantly greater depressive symptoms at early midlife.

Discussion

In this study, we examined how health insurance coverage was associated with early midlife depressive symptoms using data from Waves I, IV, and V of Add Health. We found that early midlife adults with public or no health insurance had significantly greater depressive symptoms relative to early midlife adults with private health insurance. These results align with prior research showing that Medicaid recipients tend to have poorer mental health outcomes, which may stem from differences in healthcare access, quality, and continuity of care [22, 40–42]. Similarly, being uninsured has been consistently associated with worse mental health outcomes, as the lack of healthcare coverage often leads to reduced access to preventive and ongoing care, financial barriers to treatment, and psychological distress [43–45]. Our findings also show how health insurance coverage was associated with variation in mental health outcomes across race/ethnicity at early midlife.

One of our primary findings was that the association between health insurance coverage and depressive symptoms varied across race and ethnicity, aligning with prior work [34, 35]. Structural barriers such as systemic racism, cost-sharing requirements, and competing responsibilities (e.g., employment and caregiving) may impede minoritized populations from fully benefiting from insurance coverage [46]. Even with coverage, health access challenges such as high patient loads and inadequate mental health integration in primary care may limit access to necessary mental health services. Simply having health insurance may not be enough to ensure mental health care utilization and well-being [47].

Our findings also demonstrate a unique significant association between public health insurance and depressive symptoms among Hispanic early midlife adults. More than 55% of Hispanic adults in the U.S. reported being underinsured, which includes those who lack insurance and those who have coverage but could not afford out-of-pocket costs [48]. Underinsurance also includes individuals who experience insurance churn or gaps in coverage, which can disrupt access to consistent care. Approximately 11% of U.S. adults with Medicaid reported being underinsured, and nearly 40% of Hispanic adults have reported that delaying or forgoing care negatively impacted their health [48]. These challenges have been particularly pronounced among individuals with lower incomes and those with greater health needs—populations that are disproportionately represented among Medicaid recipients [49].

We also found no significant differences in depressive symptoms between uninsured and privately insured Hispanic early midlife adults. Prior research has documented worse health outcomes among the uninsured; however, opportunities for coverage enrollment, such as emergency department-based insurance assistance, may not be systematically available or prioritized for early midlife adults without chronic conditions [50]. Additionally, enrollment barriers may be larger for individuals without immediate or perceived healthcare needs, especially when competing priorities like employment and caregiving take precedence [51]. Another factor to consider is immigration status. In our study, 21% of Hispanic adults—and 27% of Asian adults—identified as immigrants. The lack of association between being uninsured and depressive symptoms among Hispanic and Asian early midlife adults may reflect the “immigrant health paradox,” in which immigrant adults and children of immigrants often report better health outcomes compared to U.S.-born individuals, although this advantage appears to diminish over time [52].

Notably, we observed no significant differences in depressive symptoms across insurance coverage among Black and Asian adults at early midlife. These findings may reflect additional psychosocial barriers to mental health care that are experienced by all members of these groups regardless of insurance status. Black individuals experience higher levels of stigma [53–58] toward mental illness and are less likely to seek treatment for mental health challenges [53, 59, 60], including depression, compared to White individuals. Cultural contexts, across Black ethnicities, shape how individuals make meaning of mental illness and receipt of treatment [61]. In the health care setting, stigma and lacking cultural competency has been associated with increased frequency of misdiagnosis of Black people, which may translate to inappropriate treatment approaches [62, 63]. These disparities compound with decades of racialized health system misconduct to manufacture medical mistrust [64–67] and poorer mental health outcomes [68–70] among Black adults. Asian Americans have the lowest rates of mental health care utilization of any racial/ethnic group in the U.S [71]. Asian individuals may also experience stigma and shame related to mental health care [72, 73]. However, drivers of this psychosocial barrier may be differently sourced. A systematic review of mental health help-seeking among Asian Americans highlighted the model minority stereotype as a driver of minimizations of psychological distress and less favorable attitudes on help-seeking behavior [74–76]. Additional documented barriers to mental health care among Asian adults include lack of knowledge of treatment options and the mental health systems [77–79], low density of culturally and linguistically adept services and providers [80], and prohibitive costs [72].

Of note, our results also indicate that White adults with public or no health insurance reported higher levels of depression compared to their privately insured counterparts. These findings align with existing evidence that insufficient health insurance is associated with greater psychological distress among adults [81, 82]. Specifically, White adults may be particularly affected by this dynamic, given the higher overall rates of depression and mental healthcare utilization in this population compared to other racial and ethnic groups. While adults from other racial and ethnic groups often face systemic barriers to care [83–88], a limited or lack of health insurance may also have destabilizing effect for White adults due to fewer cultural expectations of healthcare inaccessibility. As White adults face downward social mobility or chronic health concerns, they may also be more likely to recognize and attempt to seek care for mental health struggles [89]. In such cases, instances of unmet need for mental health care among White adults may heighten feelings of psychological distress and contribute to greater depression symptomatology.

Another potential explanation for variation in the association between health insurance coverage and mental health across racial and ethnic groups may be the measurement of depressive symptoms. Our study used an abbreviated version of the CES-D scale, primarily capturing negative affective symptoms such as sadness and depression. Prior research suggests that self-reported feelings of sadness may be expressed differently based on cultural norms, potentially leading to measurement bias [90–93]. These variations may influence the associations observed and demonstrate the need for future research examining the applicability of depression screening tools across diverse racial and ethnic groups.

This study is not without limitations. First, its cross-sectional design limits causal interpretations. However, we leveraged the longitudinal nature of Add Health to adjust for confounders from adolescence, including health insurance status, depressive symptoms, self-rated health, and parental socioeconomic status. Second, mental health status may influence health insurance access, creating a bidirectional relationship that our analysis could not fully capture. More specifically, there is potential for reverse causality, in that individuals with poorer mental health may face greater health insurance coverage disruptions due to greater job instability [50]. Third, health insurance coverage may be influenced by self-selection – for example, healthy individuals may choose to be uninsured over enrolling in any health insurance, thus potentially being a source of bias [94, 95]. Fourth, insurance coverage was assessed at single time points at Waves I and V, preventing a detailed analysis of health insurance stability and its impact on depressive symptoms. Fifth, American Indian or Alaska Native participants (n = 295) and those who identified as some other race (n = 76) were excluded due to their small sample size in the Add Health dataset, limiting generalizability to this group. Lastly, due to missing data on the independent, dependent, and control variables in the analytic sample, there may be greater bias in the representability and generalizability of the results.

Despite these limitations, our study has several strengths. Add Health is the largest nationally representative study of adolescence and early midlife in the U.S., allowing us to provide the first nationally representative analysis of health insurance coverage and mental health at early midlife while controlling for early life course confounders. We were also able to examine racial and ethnic differences through subgroup analyses in Add Health data. In addition, by focusing on early midlife—a period marked by compounding stressors such as child-rearing, caregiving for aging parents, career advancement, and emerging health concerns [1]—our study highlighted an often-overlooked life stage in mental health research.

Conclusions

In conclusion, our study showed that disparities in health insurance access were associated with depressive symptoms at early midlife, emphasizing the need to expand access to quality healthcare across racial and ethnic groups to reduce mental health disparities. Future research can examine the mechanisms through which different types of health insurance coverage influence mental health and factors that may impact individuals’ self-selection into health insurance types, including job instability. In addition, using causal methodological approaches, such as differences-in-differences to look at health insurance policy changes across time, as well as an instrumental variables approach to help account for self-selection into health insurance types and reverse causality, could strengthen the evidence base for policy interventions aimed at expanding healthcare coverage to improve mental health outcomes. Addressing the specific healthcare needs of racially and ethnically diverse populations is necessary to ensure equitable access to healthcare and reduce mental health disparities across race and ethnicity.

Biographies

Xing Zhang

is an Assistant Professor of Population Health in the College of Health Solutions at Arizona State University. As a public policy demographer, her research focuses on the role of parent-child relationships in shaping young and early midlife adults’ health and social inequality from adolescence to early midlife, and how this varies across race, ethnicity, gender, and socioeconomic status.

Leslie B. Adams

is an Assistant Professor in the Division of Public Mental Health and Population Sciences at Stanford University School of Medicine, where she focuses on addressing mental health disparities among Black boys and men. As a behavioral scientist, her research emphasizes the role of structural racism, gender norms, and psychosocial stressors in influencing mental health outcomes. 

Tiffany L. Lemon

is an Assistant Professor of health services research at the Arizona State University’s College of Health Solutions and a faculty affiliate of the Center for Health Information and Research and the Southwest Interdisciplinary Research Center. Dr. Lemon’s current work leverages longitudinal data and causal inference methodologies to estimate the effects of health insurance coverage on patient outcomes including health care utilization, HIV disease progression, and health-related quality of life. 

Authors’ contributions

X.Z., L.B.A., and T.L.L. contributed to conceptualization, project administration, writing – original draft, and writing – review and editing. X.Z. contributed to formal analysis and methodology. All authors reviewed the manuscript.

Funding

This research uses data from Add Health, funded by grant P01 HD31921 (Harris) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), with cooperative funding from 23 other federal agencies and foundations. Add Health is currently directed by Robert (A) Hummer and funded by the National Institute on Aging cooperative agreements U01 AG071448 (Hummer) and U01AG071450 (Aiello and Hummer) at the University of North Carolina at Chapel Hill. Add Health was designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill. Dr. Leslie B. Adams was funded by the National Institute of Mental Health (K01MH127310). This study uses data from Add Health.

Data availability

The data that support the findings of this study are available from the University of North Carolina at Chapel Hill, but restrictions apply to the availability of these data and so are not publicly available. Information on how to access restricted data from the National Longitudinal Study of Adolescent to Adult Health is here: [https://addhealth.cpc.unc.edu/](https:/addhealth.cpc.unc.edu).

Declarations

Ethics approval and consent to participate

Dr. Xing Zhang received IRB approval to conduct analyses for the study (Arizona State University Study ID: 00012640). The study adhered to the Declaration of Helsinki. Add Health participants provided written informed consent for participation in all aspects of Add Health in accordance with the University of North Carolina School of Public Health Institutional Review Board guidelines that are based on the Code of Federal Regulations on the Protection of Human Subjects 45CFR46: https://www.hhs.gov/ohrp/humansubjects/guidance/45cfr46.html.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.