Immigrants Diminished Protective Effects of Educational
Attainment against Depressive Symptoms in the National
Health and Nutrition Examination Survey (NHANES
2005-2016)

Assari S; Zare H

Research Article

Austin J Public Health Epidemiol. 2021; 8(4): 1110.

Immigrants’ Diminished Protective Effects of Educational Attainment against Depressive Symptoms in the National Health and Nutrition Examination Survey (NHANES 2005-2016)

Assari S^1,2,3* and Zare H^4,5

¹Marginalization-Related Diminished Returns (MDRs) Research Center, Charles R. Drew University of Medicine and Science, California, USA

²Department of Family Medicine, Charles R. Drew University of Medicine and Science, California, USA

³Department of Urban Public Health, Charles R. Drew University of Medicine and Science, California, USA

⁴Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Maryland, USA

⁵University of Maryland Global Campus, Health Services Management, Maryland, USA

*Corresponding author: Shervin Assari, Marginalization-Related Diminished Returns (MDRs) Research Center, Charles R. Drew University of Medicine and Science, 1731 E 120^th St, Los Angeles, California, USA

Received: September 22, 2021; Accepted: October 23, 2021; Published: October 30, 2021

Abstract

Background: Socioeconomic Status (SES) indicators such as educational attainment protect people against health problems, including but not limited to depressive symptoms. However, according to the marginalization-related Diminished Returns Framework (MDRs), SES indicators such as educational attainment show weaker health effects for marginalized than for socially privileged groups. We conducted this study-built on the MDRs-with two aims: First, to test the association between educational attainment and depressive symptoms, and second, to test variation in this association by immigration status.

Methods: This cross-sectional study used the National Health and Nutrition Examination Survey (NHANES 2005-2016) data. Participants included 28,682 adults who were either non-immigrant (US-born) or immigrant. Demographic factors (age, sex, race, ethnicity, and marital status), SES (educational attainment), and depressive symptoms (Patient Health Questionnaire-9 scale) were measured. Weighted Negative Binomial Regression (NBREG) models were used in Stata to adjust for the complex sample design of the NHANES. Models without and with interaction terms were estimated in the pooled sample and by immigration status.

Results: Overall, high educational attainment showed an inverse association with depressive symptoms. However, as documented by statistical interactions between educational and immigrant status showed that immigrants with college education or above had experienced higher depressive symptoms (IRR: 1.26; CI: 1.08-1.48) than US-born individuals with college education or above.

Conclusion: Educational attainment may have a differential association with the depressive symptoms of immigrant and non-immigrant people. Immigrant people report high depressive symptoms despite their high education.

Keywords: Depressive symptoms; Education; Socioeconomic status; Population groups

Background

The protective effects of Socioeconomic Status (SES) indicators such as educational attainment are well established [1]. Scholars such as Marmot [2,3], Hayward [4-6], Link and Phelan1, Ross and Miroswky [7-9], and others [10] have provided ample theoretical and empirical evidence explaining why educational attainment is associated with better health. In addition, although educational attainment is protective against almost all health problems [1], its protective effect against depressive symptoms is also well-known [11,12].

The Marginalization-related Diminished Returns (MDRs) phenomenon [13,14] refers to weaker health effects of educational attainment for marginalized communities. However, most of this work has focused on ethnic minorities [13,14]. As shown by Assari [13,14], Ferarro [15], Thorpe [16-18], Hudson [19-21], and others [22], weaker effects of educational attainment on health are found for ethnic minorities than for ethnic majorities. Other investigators such as Kaufman [23], Braveman [24], Shapiro [25,26], Williams [27,28], Ceci [29], and Navarro [30-32] have shown that educational attainment and other SES indicators may not even be comparable across social groups; as a result, the health effects of such SES indicators may also be unequal across various social groups [13,33]. For example, Ceci highlighted the differences between the Haves and the Have-Nots in their capacity to uptake any available or new resource (e.g., SES indicators) [29]. He argues that when resources become available, the Have-Nots (socially marginalized people) may be at a relative disadvantage for turning those resources into outcomes [29]. While any marginalized group may show MDRs (diminished health effects of SES), these MDRs are most robust for comparing ethnic groups, and they are best described for Blacks (relative to Whites) [13,14].

While these MDRs hold across SES indicators and health outcomes, they are best described for ethnic minorities than for immigrants. A literature review shows only five studies on MDRs of SES indicators in immigrant people, and such literature is scattered across age groups of children [34], adults [35], and older adults [36] and various SES indicators such as income [36] and other SES indicators [34]. For example, in one study, income showed a more significant effect on the mental health of US-born than on immigrant older adults [36]. However, these are also rarely shown for psychological well-being [35], such as depressive symptoms [34]. As such, it is important to study the MDRs of education on depressive symptoms for immigrants. This is important because MDRs may be more robust for more distant (e.g., education) than proximal social determinants (e.g., income or employment). This is probably because more social processes can hinder the effects on health of educational attainment than other SES indicators such as income and employment on health [37]. Educational attainment may not result in the same employment for diverse people because of the labor market discrimination [38]. As such, we may expect stronger MDRs for educational attainment than for other SES indicators such as income and employment. As some of the MDRs of educational attainment may be due to differential employment opportunities or income, controlling for employment and income may reduce the significance of MDRs due to educational attainment [37].

The MDRs framework [13,14] can be regarded as a paradigm shift in health disparities research because MDRs perform the following: (a) seek how economic and health effects of available SES indicators vary across subpopulations, (b) explore health disparities across the full SES spectrum, (c) allow SES returns to across groups, (d) use a moderated-mediation rather than a mediation model, (e) test the non-linear and non-additive effects of group membership and SES which are more realistic than universal average effects, and (e) explain why some health gaps may widen rather than narrow as SES increases [13,14].

Aims

In response to the gap in the literature, we conducted a secondary analysis of the National Health and Nutrition Examination Survey (NHANES) to determine the association between educational attainment and depressive symptoms by immigration status. While we hypothesize an inverse association between educational attainment and depressive symptoms, in line with the MDRs framework, we hypothesized that a weaker association between educational attainment and depressive symptoms would be weaker for immigrants than for US-born adults. As a result, we expected high depressive symptoms in immigrant people across educational attainment levels. In contrast, we expected low depressive symptoms for highly educated US-born people.

Materials and Methods

We used the National Health and Nutrition Examination Survey (NHANES) data between 2005-2016, [39]. The NHANES is a crosssectional survey that provides nationally representative health and nutritional status estimates for the US population. The response rate for this data between 1999-2016 reported 73.2% [40,41]. For this analysis, we included 28,682 individuals who were 20 years old and older. NHANES used a nationally representative sampling. The sampling strategy was clustered and stratified.

Outcome Variable

Depressive symptoms were defined using the Patient Health Questionnaire-9 cutoff of 10 or higher. The PHQ-9 is a “clinically validated survey with a sensitivity of 88% and a specificity of 88% at a cutoff score of 10 or higher” [42]. This measure asks people to rate how often over the past two weeks they experienced depressive symptoms, such as restless sleep, poor appetite, and feeling lonely. Each item was scored on a 4-point ordinal scale for frequency (0, not at all; 1, several days; 2, more than half the days; 3, nearly every day). The total score was calculated by finding the sum of 9 items; this approach yielded a maximum score of 27 [43]. We used this scale as the dependent variable in all NBREG models.

Main Independent Variables

The main independent variables of interest were educational attainment levels. Educational attainment was defined as a categorical variable (less than high school graduate, high school graduate or general equivalency diploma, some college or AA degree, college graduate or above).

Covariate

Covariates included race, ethnicity, age (years), sex, and marital status (1 = married, 0 = otherwise). Race was a nominal variable (non-Hispanic White = 0, non-Hispanic Black =1, Hispanic = 3, and other = 4).

Analytical Strategy

We used Stata statistical software version 15 to perform all analyses. We used descriptive analysis to compare the mean and proportional differences between immigrant and non-immigrant people for all study variables. Demographics, SES, and depressive symptoms were evaluated using unequal variances t-tests and chisquare. We conducted several sets of weighted negative binomial regression models [44]. From our regression models, we reported Incidence-Rate Ratios (IRR) and the corresponding 95% Confidence Intervals (CI) [45,46]. For the first set of analyses, we ran additive models. Then, to find the impact of educational attainment interaction on depressive symptoms, we ran a model with an interaction between immigration status and educational attainment. Finally, for the last set of analyses and because the interaction between immigration status and educational attainment was significant (p < 0.001), we stratified the analyses by immigration status. All analyses were weighted using the NHANES individual-level sampling weights for 2005-2016 (6 waves of data). As such, the estimates are representative at the national level for the US civilian population [47].

Results

Descriptive Data

A total of 28,682 individuals entered our analysis. From all participants, 70.2% (n =16,624) were non-Hispanic White (NHW), 10.9% (n =2,581) non-Hispanic Black (NHB), 12.9% (n =3,055) Hispanics and 3.1% (n = 1,421) other racial groups. The prevalence of individuals with depressive symptoms was 7.7%. The mean age of the participants was about 47.5 years (SD =14.4); 50.6% were female and 63.3% were married. Of all, 59.0% had an educational attainment higher than a high school degree (Table 1).

Table 1: Demographic Characteristics of Nationally Representative Samples of US-Born and Immigrants Aged 20 and above, NHANES 2005-2016.




  
    
    Immigrant    (n = 5,805)
    Non-Immigrant    (n = 17,877)
    All (n    = 23,682)
    p-value
  
  
    Depressive Symptoms
    Mean/%
    (SD)
    Mean/%
    (SD)
    Mean/%
    (SD)
    
  
  
    PHQ-D Score
    7
    (28.2)
    7.9
    -21.6
    7.7
    (22.7)
    0.093
  
  
    If PHQ>10a
    2.8
    (4.4)
    3.0
    (3.3)
    3.0
    (3.5)
    0.000
  
  
    PHQ Categories
    
    
    
    
    
    
  
  
    Minimal
    79.1
    (45.1)
    77.2
    (33.7)
    77.4
    (35.6)
    0.131
  
  
    Mild
    14
    (38.4)
    15
    (28.6)
    14.8
    (30.3)
    
  
  
    Moderate
    4.4
    (22.6)
    4.9
    (17.3)
    4.8
    (18.3)
    
  
  
    Moderate-Severe and Severe
    2.6
    (17.6)
    2.9
    (13.6)
    2.9
    (14.3)
    
  
  
    Immigrant
    100
    
    100
    
    14.5
    (30.0)
    
  
  
    Socio-demographic
    
    
    
    
    
    
  
  
    Age (years)
    44.3
    (16.9)
    48
    (13.7)
    47.5
    (14.4)
    0.000
  
  
    Female
    46.8
    (55.3)
    51.2
    (40.1)
    50.6
    (42.6)
    0.000
  
  
    Married
    69.3
    (51.2)
    62.3
    (38.9)
    63.3
    (41.0)
    
  
  
    Educational attainment
    
    
    
    
    
  
  
    Less than high school
    36
    (53.2)
    13.8
    (27.7)
    17.0
    (32.0)
    0.000
  
  
    High school graduate/GED
    17.1
    (41.7)
    24.2
    (34.3)
    23.1
    (35.9)
    
  
  
    Some college or AA degree
    20.8
    (45.0)
    33.3
    (37.8)
    31.5
    (39.5)
    
  
  
    College graduate or above
    26.1
    (48.7)
    28.7
    (36.3)
    28.3
    (38.4)
    
  
  
    Race/Ethnicity
    
    
    
    
    
    
  
  
    White Non-Hispanic
    16.5
    (41.2)
    79.3
    (32.5)
    70.2
    (38.9)
    0.000
  
  
    Black Non-Hispanic
    6.3
    (26.9)
    11.7
    (25.8)
    10.9
    (26.6)
    
  
  
    Hispanics
    54.2
    (55.2)
    5.8
    (18.8)
    12.9
    (28.5)
    
  
  
    Other
    22.9
    (46.6)
    3.1
    (14.0)
    6.0
    (20.2)
    
  
  
    Note: Depressive symptoms    categories were calculated using the Patient Health Questionnaire–9: none    (0-4), mild (5-9), moderate (10-14), moderately severe (15-19), and severe    (20). For this table, we combined moderately severe and severe as one group.    The Percentages were weighted to the population of noninstitutionalized US    adults aged 20 years or older. aWe created a dummy variable if the    score was equal to or higher than 10.



Table 1: Demographic Characteristics of Nationally Representative Samples of US-Born and Immigrants Aged 20 and above, NHANES 2005-2016.

Bivariate analysis

As Table 1 shows, immigrant participants were younger than non-immigrant participants by four years. Education was also higher in non-immigrant than immigrant participants. While immigrants were more likely to be Hispanic, non-immigrants were more likely to be NHW and NHB. Immigrants and non-immigrants also varied in sex and likelihood of being married. Compared to non-immigrant people, immigrant individuals had lower probability of having PHQ- 9 scales higher than 10 than US-Born, 2.8 (SD = 4.4) vs. 3.0 (SD = 3.3) (See Table 1). (Figure 1-Figures 4 and Table 2) show the distribution of PHQ-9 scales between immigrants and US-born.

Table 2: Prevalence of Depressive Symptoms in US-Born and Immigrants Aged 20 and above, NHANES 2005-2016.




  
    PHQ Scale
    Less    than

      high    school
    High    school

      graduate/GED
    Some    college

      or AA    degree
    College    graduate

      or    above
  
  
    Mean/%
    (SD)
    Mean/%
    (SD)
    Mean/%
    (SD)
    Mean/%
    (SD)
  
  
    Immigrant
    3.24
    0.11
    2.83
    0.19
    2.76
    0.14
    2.07
    0.11
  
  
    US-Born
    4.13
    0.15
    3.32
    0.1
    3.16
    0.8
    2.12
    0.5
  
  
    Overall
    3.85
    0.11
    3.26
    0.09
    3.11
    0.08
    2.11
    0.04
  
  
    p-value (Immigrant vs. US-Born
    0
    
    0
    
    0
    
    0
    
  
  
    PHQ Category
    
    
    
    
    
    
    
  
  
    Immigrant
    
    
    
    
    
    
    
    
  
  
    Minimal
    1,842
    72.5
    752
    78.2
    860
    78.6
    1,036
    85.5
  
  
    Mild
    411
    16.2
    142
    14.8
    158
    14.4
    129
    10.7
  
  
    Moderate
    174
    6.9
    37
    3.8
    48
    4.4
    32
    2.6
  
  
    Moderate-Severe and severe
    112
    4.4
    31
    3.2
    28
    2.6
    14
    1.2
  
  
    Total Immigrant
    2,539
    
    962
    
    1,094
    
    1,211
    
  
  
    US-Born
    
    
    
    
    
    
    
    
  
  
    Minimal
    2,385
    67.2
    3,332
    73.4
    4,289
    74.3
    3,405
    84.7
  
  
    Mild
    629
    17.7
    754
    16.6
    948
    16.4
    460
    11.4
  
  
    Moderate
    316
    8.9
    284
    6.3
    334
    5.8
    101
    2.5
  
  
    Moderate-Severe and severe
    220
    6.2
    170
    3.7
    198
    3.4
    53
    1.3
  
  
    Total US-Born
    3,550
    
    4,540
    
    5,769
    
    4,019



Table 2: Prevalence of Depressive Symptoms in US-Born and Immigrants Aged 20 and above, NHANES 2005-2016.

Figure 1:

    
    
    Figure 1:

Figure 2:

    
    
    Figure 2:

Figure 3:

    
    
    Figure 3:

Figure 4:

    
    
    Figure 4:

Pooled sample models

Table 3 shows the regression models in the pooled sample. According to the model without interaction term (Model 1), high educational attainment was inversely associated with depressive symptoms. However, according to the models with interaction terms (Model 2), immigrants with college education or above experienced a higher relative ratio (IRR; C: 1.26; [1.08-1.48]) than US-born with college education or above.

Table 3: Weighted Negative Binomial Regression Estimates in US adults aged 20 and above, NHANES 2005-2016.




  
    
    Model 1 IRR [95% CI]
    Model 2 IRR [95% CI]
  
  
    Age (years)
    1.00*** [1.00-1.00]
    1.00** [1.00-1.00]
  
  
    Female
    1.38*** [1.32-1.44]
    1.38*** [1.32-1.44]
  
  
    Married
    0.76*** [0.72-0.79]
    0.76*** [0.72-0.79]
  
  
    Education (Ref. Less than high school)
  
  
    High school graduate/GED or equivalent
    0.84*** [0.79-0.90]
    0.82*** [0.76-0.89]
  
  
    Some college or AA degree
    0.77*** [0.71-0.83]
    0.75*** [0.69-0.82]
  
  
    College graduate or above
    0.55*** [0.51-0.59]
    0.53*** [0.48-0.57]
  
  
    Immigration status
    
  
  
    Non-US-Born (immigrant)
    0.85*** [0.79-0.92]
    0.76*** [0.68-0.84]
  
  
    Race/Ethnicity (Ref. NHW)
  
  
    Black Non-Hispanic
    0.99 [0.93-1.06]
    0.99 [0.93-1.05]
  
  
    Hispanics
    1.04 [0.97-1.12]
    1.07 [0.99-1.15]
  
  
    Other
    1.12* [1.02-1.22]
    1.10* [1.01-1.20]
  
  
    Education # Immigration status
  
  
    Immigrant with high school graduate/GED or    equivale
    NA
    1.08 [0.92-1.27]
  
  
    Immigrant with some college or AA degree
    NA
    1.12 [0.98-1.28]
  
  
    Immigrant with college education or above
    NA
    1.26** [1.08-1.48]
  
  
    N
    23,682
    23,682
  
  
    *p<0.05, **p<0.01, ***p<0.001

      IRR: Incidence Rate Ratios.



Table 3: Weighted Negative Binomial Regression Estimates in US adults aged 20 and above, NHANES 2005-2016.

Stratified models

Table 4 shows the weighted negative binomial regression estimates for immigrants and US-born. Education is protective for all levels of education in US-born, but only in immigrants with college education or above (IRR: 0.73; CI: 0.64-0.83). Sex and marital status are predictors for immigrants and US-born.

Download PDF

Citation: Assari S and Zare H. Immigrants’ Diminished Protective Effects of Educational Attainment against Depressive Symptoms in the National Health and Nutrition Examination Survey (NHANES 2005-2016). Austin J Public Health Epidemiol. 2021; 8(4): 1110.

Home

Journal Scope

Editorial Board

Instruction for Authors

Submit Your Article

	Model 1 IRR [95% CI]	Model 2 IRR [95% CI]
Age (years)	1 [1.00-1.00]	1.00*** [1.00-1.00]
Female	1.39*** [1.28-1.51]	1.38*** [1.31-1.44]
Married	0.71*** [0.64-0.79]	0.77*** [0.73-0.80]
Education (Ref. Less than high school)
High school graduate/GED or equivalent	0.92 [0.80-1.06]	0.82*** [0.76-0.89]
Some college or AA degree	0.89 [0.78-1.01]	0.74*** [0.68-0.81]
College graduate or above	0.73*** [0.64-0.83]	0.52*** [0.48-0.57]
Race/Ethnicity (Ref. NH White)
NH Black	0.93 [0.78-1.12]	0.99 [0.93-1.06]
Hispanics	1.14 [0.99-1.30]	1.01 [0.93-1.09]
Other	0.97 [0.81-1.17]	1.23*** [1.10-1.38]
N	5,805	17,877
p<0.05, p<0.01, **p<0.001 IRR: Incidence Rate Ratios.