Abstract
Objective
This study investigated the association of race/ethnicity, dietary intake and physical activity with depression and potential other barriers associated with the use of mental health services among depressed people.
Methods
We used the nationally representative data, 2011–2016 National Health and Nutrition Examination Survey. Depression status was defined using a Patient Health Questionnaire. Multivariable logistic regressions were conducted on depression status and the use of mental health specialists among depressed adults, accounting for the complex sampling design.
Results
The prevalence of depression was 8.3% with substantial racial/ethnic differences (8.0% for White, 3.1% for Asian, 9.2% for Black, 7.6% for Mexican Hispanics, 13.0% for Other Hispanics). Good/acceptable diet and a high level of physical activity was negatively associated with depression. Among depressed people, no significant racial/ethnic differences were observed in using mental health specialists.
Conclusion
Prevalence for depression was lower among people who have good or acceptable diet and moderate physical activity. These modifiable factors as well as race/ethnicity should be incorporated into psychotherapeutic interventions to improve depression.
Keywords: race/ethnicity, depression, nutrition, exercise, health eating index, mental health utilization
Background
Depression is the most common mental disorder in the world [1]. The National Institute of Mental Health estimated that 17.3 million adults in the United States (US) had at least one major depressive episode [2]. The lifetime prevalence of depression is about 7% with higher rates among women than men [3]. Depression is the leading cause of disability [4] and is known to be associated with lowered quality of life, elevated medical costs, and increased risk for suicide and mortality [5].
Factors associated with depression have been investigated in various sociodemographic, medical and behavioral variables. A meta-analysis showed that sex, family and personal history of mental disorder, and current substance misuse are the significant risk factors for depression [6]. As for behavioral variables, there has been supporting evidence that physical activity and diet quality or pattern are associated with depression [7–10]. For example, a systematic review studying 6,363 papers showed the protective effect of physical activity on depression [10]. A meta-analysis study confirmed that adherence to a healthy dietary pattern, comprising higher intakes of fruit and vegetables, fish and whole grains, was associated with a reduced likelihood of depression in adults [7]. As for physical activity, De Mello et al. reported that people who do not engage in physical activity were more likely to show depression [8].
Race/ethnicity has also received attention in explaining the difference in the prevalence of depression. For example, Riolo et al. showed the prevalence of depressive disorder was significantly higher in Whites than in Blacks and Mexican Americans [11] and Steffens et al. showed a lower prevalence rate in Blacks compared to Whites [12]. Regarding Asians, previous observational studies revealed lower rates of depression compared to other racial/ethnic groups [13–15]. However, some other studies reported higher prevalence rates of depression among minority groups including Asians, compared to Whites [16]. Because of these contradictory results, it is still worthwhile to investigate racial/ethnic differences including Asians. In addition, Asians are often disregarded or merged into other racial/ethnic group because of their relatively small population size in the US. However, since Asians are one of the fastest-growing subpopulations in the country [17], empirical investigations are needed to explore whether there is a difference in prevalence of depression and to identify important variables associated with depression for this ethnic group. Because of the small population size, only a few published articles exist exploring the factors associated with depression by race/ethnicity including Asians.
Although many people have been affected by depressive symptom during their lifetime, only less than half of the people with depression stated that they have received treatment for it [18]. Like prevalence, disparities were also reported in the utilization of mental health service among people with depression. Studies found that racial/ethnic minorities are less likely to use mental health care than Whites [19–23]. However, most research studies have addressed disparities in mental health utilization among Blacks or Hispanics with depression. Few studies have included Asians as a separate group in their analyses [19, 23].
To our knowledge, a few studies focused on racial/ethnic difference in depression and mental health utilization with Asians, and included modifiable variables such as dietary intake and physical activity. This current study was aimed to examine the effect of race/ethnicity, dietary intake, and physical activity on depression adjusting for other potential factors employing a nationally representative data; to investigate the effect of dietary intake and physical activity as well as other variables by race/ethnicity; and to identify variables associated with the use of mental health specialists among depressed people.
Methods
Data Source and Study Sample
The data from the National Health and Nutrition Examination Survey (NHANES) 2011–2016 were utilized in this study. The NHANES, a nationally representative sample with complex multistage sampling design, biannually collects data to assess the health and nutritional status of adults and children in the US. It combines interviews and physical examinations that the participants go through at-home and clinic examinations in a Mobile Examination Center (MEC). Dietary intake data are obtained to estimate the types and amounts of foods and beverages consumed during the 24-hour period prior to the interview. The first dietary recall interview is collected in-person in the MEC and the second interview is collected by telephone 3 to 10 days later. Beginning with the 2011–2012 cycle, the NHANES over-sampled Asian Americans and provided a new race variable (RIDRETH3) to identify Asian Americans. This study includes 14,865 participants aged 20 years or older who completed the 24-hour dietary recall interviews.
Variables
Depression Status and Use of Mental Health Specialists:
Depression status was defined using the validated Patient Health Questionnaire-9 (PHQ-9) [24]. The PHQ-9 is a nine-item instrument to assess the overall impairment of depression symptoms over the past two weeks, with a total score ranging from 0 to 27. Its reliability was good in this study (Cronbach’s alpha=0.84). Depression status was clinically diagnosed if the PHQ-9 score is greater than or equal to 10 [24]. The use of mental health specialists was defined by the question ‘have you seen a mental health professional in the past year’ and coded as Yes (1)/No (0).
Race/ethnicity:
Race/ethnicity was categorized into White, Asian, Black, Mexican Hispanic, Other Hispanic, and Other.
Diet:
The Healthy Eating Index (HEI) is a measure of diet quality, independent of quantity, that can be used to assess compliance with the U.S. Dietary Guidelines for Americans and monitor changes in dietary patterns [25]. The HEI score was computed using the version 2015 scoring algorithm provided by the National Cancer Institute [26]. The HEI total score was computed based on 13 components of total vegetable, greens and beans, total fruit, whole fruit, whole grains, dairy, total protein foods, seafood and plant proteins, fatty acids, sodium, refined grains, saturated fats, and added sugars [27]. The score ranges from 0 to 100 (categorized into A: >90; B: 80–90; C: 70–80; D: 60–70; E: 50–60; F: <50). We further classified them into three groups: acceptable (A/B), needs improvement (C/D/E), and poor (F). In addition, we included caffeine intake (categorized into 0 mg, >0–200 mg, >200–400 mg, and >400 mg) and self-rated dietary pattern (categorized into excellent/very good, good, and fair/poor).
Physical Activity:
The NHANES physical activity questionnaire is based on the Global Physical Activity Questionnaire [28, 29]. The average number of hours per week spent in each activity was multiplied by the suggested metabolic equivalent of task (MET) scores (8 for vigorous physical activity and 4 for moderate physical activity) to get an estimate of MET-hours per week. MET-hours per week of vigorous and moderate leisure-time physical activity were summed to obtain an estimate of total leisure-time physical activity. Leisure-time physical activity level was then categorized as none, low (MET: >1 to 12), moderate (MET: >12 to 32), and high (MET: >32). The cut-points for low, moderate, and high leisure-time physical activity were defined by tertiles among the participants who did at least 10 minutes moderate or vigorous physical activity in a week.
Other Variables:
This study considers the following other demographic variables: age, sex, marital status, education, insurance type, and income. Age was categorized into three groups: 20–44, 45–64, and ≥65 years. Marital status was categorized into three groups as married/cohabitated, single, and other (widowed/divorced/separated). Education was categorized as less than high school, high school graduate, some college, and college graduate or more. Income was defined using the ratio of family income to poverty [FIP], categorized as poor (FIP≤1), low (1<FIP≤2), moderate (2<FIP≤4) and high (FIP>4). Insurance was categorized as no insurance, public (Medicare or Medicaid), private, and other.
As health-related risk factors, we included history of sleep disorder, number of chronic conditions, general health status (excellent/very good, good, fair/poor), and obesity status (normal or underweight: BMI<25, overweight: 25≤BMI<30, obese: BMI≥30). Chronic conditions considered included: asthma, arthritis, heart failure, coronary heart disease, angina, diabetes, heart attack, stroke, emphysema, thyroid problem, chronic bronchitis, liver problem, and cancer or malignancy. The number of these chronic conditions was further categorized into 0, 1, and 2 or more. For other behavioral risk factors, we included alcohol consumption (>2 drinks/day), and smoking (never, previous, current) as behavioral risk factors.
Statistical Analysis
Descriptive statistics were reported using unweighted frequencies with weighted percentages and weighted means together with standard errors. Bivariate association between each variable and depression status and race/ethnicity was explored using Rao-Scott’s chi-square tests for categorical variables and two sample t-tests for continuous variables. We used univariable and multivariable logistic regressions on depression. Multicollinearity was assessed by variance inflation factor. To assess the effect of nutrition and physical activity in different racial/ethnic groups, we also conducted a subgroup analysis by race/ethnicity. Then we conducted a multivariable logistic regression on use of mental health specialists among depressed people with the same variables used for depression status. A forest plot was created for a pictorial illustration of the result only for race/ethnicity and significant variables. All analyses were conducted in R version 3.5.1 using the survey package adjusting for the NHANES complex sampling design. A dietary 24-hour recall interview day weight (WTDRD1), divided by three (because of three-cycle data), was used in the analyses to get the estimates representative to the US adult population. P-value less than 0.05 was considered statistically significant.
Results
Table 1 presents participants’ characteristics. About half were women (51.6%); 24.1% were 65 years or older. The racial/ethnic distribution was 65.5% White, 5.5% Asian, 11.3% Black, 8.7% Mexican Hispanic, 6.0% Other Hispanic, and 3.0% Other. About 22% had two or more chronic conditions and 29% have a history of sleep disorder. About 26% rated their dietary pattern being fair or poor and 66.8% received a poor grade in HEI. Only 14.4% of people had a high level of total leisure-time physical activity, while 45.1% did not do a physical activity in their leisure time.
Table 1.
Descriptive Statistics and Bivariate Association with Depression Status
| Variable | Total N (W%) | Depression Status | P-value | |
|---|---|---|---|---|
| Not depressed | Depressed | |||
| (91.7%) (W%) | (8.3%) (W%) | |||
| Main Variable | ||||
| Race/Ethnicity | <0.001 | |||
| White | 5,786 (65.5) | 92.0 | 8.0 | |
| Asian | 1,656 (5.5) | 96.9 | 3.1 | |
| Black | 3,343 (11.3) | 90.8 | 9.2 | |
| Mexican Hispanic | 2,010 (8.7) | 92.4 | 7.6 | |
| Other Hispanic | 1,590 (6.0) | 87.0 | 13.0 | |
| Other | 480 (3.0) | 85.0 | 15.0 | |
| Self-rated dietary pattern | <0.001 | |||
| Excellent/very good | 4,288 (31.3) | 95.8 | 4.2 | |
| Good | 6,207 (42.3) | 93.3 | 6.7 | |
| Fair/poor | 4,364 (26.3) | 84.3 | 15.7 | |
| HEI Total | <0.001 | |||
| Poor | 9,668 (66.8) | 90.4 | 9.6 | |
| Need improvement | 4,380 (30.0) | 93.9 | 6.1 | |
| Acceptable | 517 (3.2) | 97.7 | 2.3 | |
| Caffeine | 0.067 | |||
| None | 2,324 (12.7) | 90.8 | 9.2 | |
| 1–200 mg | 8,815 (55.8) | 92.1 | 7.9 | |
| >200–400 mg | 2,646 (21.8) | 92.4 | 7.6 | |
| >400 mg | 1,080 (9.8) | 89.0 | 11.0 | |
| Total Leisure-Time Physical Activity (MET-hours/week) | <0.001 | |||
| None | 7,482 (45.1) | 88.1 | 11.9 | |
| Low (≤12) | 2,943 (20.8) | 92.7 | 7.3 | |
| Moderate (>12 to 32) | 2,547 (19.7) | 95.1 | 4.9 | |
| High (>32) | 1,872 (14.4) | 96.6 | 3.4 | |
| Socio-Demographic | ||||
| Sex | <0.001 | |||
| Female | 7,642 (51.6) | 89.8 | 10.2 | |
| Male | 7,223 (48.4) | 93.6 | 6.4 | |
| Age | <0.001 | |||
| 20–44 | 6,426 (42.0) | 92.0 | 8.0 | |
| 45–64 | 5,091 (33.9) | 90.1 | 9.9 | |
| ≥65 | 3,348 (24.1) | 93.9 | 6.1 | |
| Education | <0.001 | |||
| <High school | 3,280 (14.9) | 86.5 | 13.5 | |
| High school | 3,257 (21.1) | 90.0 | 10.0 | |
| Some college | 4,533 (33.0) | 90.9 | 9.1 | |
| College | 3,787 (31.0) | 96.0 | 4.0 | |
| Marital Status | <0.001 | |||
| Married/cohabitating | 8,723 (61.8) | 94.0 | 6.0 | |
| Never married | 2,929 (19.5) | 89.7 | 10.3 | |
| Other (divorced, separated, widowed) | 3,208 (18.7) | 86.1 | 13.9 | |
| Ratio of Family Income to Poverty | <0.001 | |||
| Poor | 3,119 (14.9) | 83.3 | 16.7 | |
| Low | 3,601 (19.8) | 88.0 | 12.0 | |
| Moderate | 3,513 (25.9) | 92.8 | 7.2 | |
| High | 3,424 (32.8) | 96.5 | 3.5 | |
| Missing | 1,208 (6.5) | 92.2 | 7.8 | |
| Health Insurance | <0.001 | |||
| No insurance | 3,065 (17.1) | 89.1 | 10.9 | |
| Public | 3,045 (15.1) | 82.5 | 17.5 | |
| Private | 7,592 (60.7) | 95.3 | 4.7 | |
| Other | 1,144 (7.2) | 85.5 | 14.5 | |
| Health-related | ||||
| # of Chronic Conditions | <0.001 | |||
| 0 | 7,030 (48.2) | 94.9 | 5.1 | |
| 1 | 3,837 (26.7) | 92.2 | 7.8 | |
| 2+ | 3,764 (25.2) | 85.8 | 14.2 | |
| General Health | <0.001 | |||
| Excellent/very good | 5,759 (46.5) | 97.7 | 2.3 | |
| Good | 5,557 (35.5) | 92.0 | 8.0 | |
| Fair/poor | 3,534 (18.0) | 75.5 | 24.5 | |
| Behavioral-related | ||||
| Alcohol | 0.027 | |||
| No | 9,462 (63.3) | 92.3 | 7.7 | |
| Yes | 4,577 (36.7) | 90.7 | 9.3 | |
| Smoking | <0.001 | |||
| Never | 8,443 (55.9) | 94.4 | 5.6 | |
| Previous | 3,469 (24.8) | 92.8 | 7.2 | |
| Current | 2,938 (19.3) | 82.6 | 17.4 | |
| Sleep disorder | <0.001 | |||
| No | 10,917 (71.0) | 95.7 | 4.3 | |
| Yes | 3,946 (29.0) | 82.1 | 17.9 | |
| Obesity | <0.001 | |||
| Normal or underweight | 4,309 (29.6) | 92.8 | 7.2 | |
| Overweight | 4,731 (33.0) | 93.8 | 6.2 | |
| Obese | 5,680 (37.4) | 89.1 | 10.9 | |
| Outcome | ||||
| Use of mental professional | <0.001 | |||
| No | 13,623 (90.8) | 93.5 | 6.5 | |
| Yes | 1,238 (9.2) | 73.5 | 26.5 | |
HEI = Health Eating Index. MET = Metabolic Equivalent of Task. W% = Weighted percentage (weighted row percentage for depression). Adjusting for the NHANES complex sampling design, Rao-Scott’s chi-square test was used to explore bivariate association with depression.
The prevalence rate of depression was 8.3%. All variables were significantly associated with depression except caffeine intake (Table 1). The depression rate for women (10.2%) was higher than that of men (6.4%). Compared to White (8.0%), Asian (3.1%) has a lower rate but other Hispanic (13.0%) has a higher rate for depression. Self-rated dietary pattern was associated with depression with the lowest (4.2%) among the people who rated their dietary pattern as excellent/very good. People whose HEI total was acceptable showed the lowest depression rate (2.3%) but people whose HEI total was poor showed the highest rate (9.6%). Depression rate was also decreased with an increased level of physical activity from 11.9% (no exercise) to 3.4% (high level exercise). Only 26.5% of people with depression visited mental health professionals in the past year.
In addition, we explored which dietary components were associated with depression (see Appendix 1). The association varied by dietary component. Compared to non-depressed people, depressed people had more unhealthy diet patterns. For example, their diet intake was lower in vegetables, fruit, and whole grains but higher in sodium than their counterparts.
The results of a logistic regression for depression are shown in Table 2. In the univariable regression, all variables were significantly associated with depression except caffeine intake. In the multivariable model, race/ethnicity, diet, and physical activity were significantly associated with depression. For race/ethnicity, Asians, Blacks, and Mexican Hispanics were less likely to be depressed than Whites, ranging 0.54–0.66 in odds ratio [OR]. For self-rated dietary pattern, people who reported they have poor/fair dietary pattern showed a 46% higher depression rate than people who reported they have good dietary pattern. People who received an acceptable grade in HEI showed a 55% lower depression rate than those who received a poor grade. For physical activity, people engaging a high level of physical activity in leisure time showed a 27% lower depression rate than people who do not exercise.
Table 2.
Univariable and Multivariable Logistic Regressions on Depression
| Variable | Univariable Logistic Regression OR (95% CI) | Multivariable Logistic Regression OR (95% CI) |
|---|---|---|
| Main Variable | ||
| Race/Ethnicity (Ref: White) | ||
| Asian | 0.37 (0.26–0.53)*** | 0.59 (0.39–0.90)* |
| Black | 1.17 (0.94–1.45) | 0.66 (0.52–0.84)* |
| Mexican Hispanic | 0.95 (0.69–1.31) | 0.54 (0.38–0.78)* |
| Other Hispanic | 1.73 (1.29–2.33)*** | 1.26 (0.90–1.76) |
| Other | 2.04 (1.33–3.11)** | 1.32 (0.81–2.15) |
| Self-rated dietary pattern (Ref: Excellent/very good) | ||
| Good | 1.63 (1.23–2.16)*** | 0.99 (0.72–1.36) |
| Fair/poor | 4.22 (3.20–5.57)*** | 1.46 (1.03–2.07)* |
| HEI Total (Ref: Poor) | ||
| Need Improvement | 0.61 (0.49–0.75)*** | 0.94 (0.75–1.17) |
| Acceptable | 0.22 (0.13–0.37)*** | 0.45 (0.26–0.80)* |
| Caffeine (Ref: None) | ||
| 1–200 mg | 0.85 (0.63–1.15) | 0.88 (0.61–1.25) |
| >200–400 mg | 0.82 (0.58–1.16) | 0.88 (0.60–1.30) |
| >400 mg | 1.22 (0.84–1.78) | 0.94 (0.59–1.49) |
| Total Leisure-Time Physical Activity (Ref: None) | ||
| Low | 0.58 (0.45–0.74)*** | 0.91 (0.69–1.19) |
| Moderate | 0.38 (0.28–0.51)*** | 0.79 (0.57–1.10) |
| High | 0.26 (0.19–0.37)*** | 0.63 (0.43–0.90)* |
| Socio-Demographic | ||
| Sex (Ref: Female) | ||
| Male | 0.60 (0.50–0.72)*** | 0.63 (0.53–0.76)** |
| Age (Ref: 20–44 yrs.) | ||
| 45–64 yrs. | 1.26 (1.04–1.52)* | 0.95 (0.75–1.21) |
| ≥65 yrs. | 0.75 (0.59–0.95)* | 0.55 (0.39–0.80)* |
| Education (Ref: <High school) | ||
| High school graduate | 0.71 (0.58–0.87)** | 0.90 (0.70–1.15) |
| Some college | 0.64 (0.50–0.81)*** | 0.97 (0.73–1.29) |
| College | 0.27 (0.19–0.38)*** | 0.86 (0.56–1.31) |
| Marital Status (Ref: Married/cohabitating) | ||
| Never married | 1.81 (1.47–2.24)*** | 1.68 (1.30–2.16)** |
| Other | 2.55 (2.16–3.02)*** | 1.45 (1.17–1.80)* |
| Income (Ref: Poor) | ||
| Low | 0.68 (0.55–0.84)*** | 0.92 (0.72–1.16) |
| Moderate | 0.38 (0.30–0.49)*** | 0.83 (0.64–1.07) |
| High | 0.18 (0.13–0.25)*** | 0.56 (0.38–0.84)* |
| Missing | 0.42 (0.32–0.57)*** | 0.75 (0.50–1.13) |
| Health Insurance (Ref: No insurance) | ||
| Public | 1.75 (1.42–2.14)*** | 1.24 (0.94–1.63) |
| Private | 0.40 (0.31–0.51)*** | 0.67 (0.52–0.87)* |
| Other | 1.39 (0.98–1.96)+ | 1.42 (0.94–2.14) |
| Health-related | ||
| # of Chronic Conditions (Ref: 0) | ||
| 1 | 1.56 (1.26–1.94)*** | 1.00 (0.77–1.29) |
| 2+ | 3.08 (2.60–3.65)*** | 1.32 (1.03–1.69)* |
| General Health (Ref: Excellent/very good) | ||
| Good | 3.65 (2.67–4.99)*** | 2.27 (1.62–3.18)** |
| Fair/poor | 13.57 (10.59–17.37)*** | 5.43 (3.93–7.49)*** |
| Behavioral-related | ||
| Alcohol | 1.23 (1.03–1.46)* | 1.08 (0.88–1.34) |
| Smoking (Ref: Never) | ||
| Previous | 1.31 (1.02–1.67)* | 1.00 (0.75–1.33) |
| Current | 3.52 (2.84–4.37)*** | 1.55 (1.19–2.02)* |
| Sleep disorder | 4.83 (4.16–5.61)*** | 3.20 (2.65–3.87)*** |
| Obesity (Ref: Normal/underweight) | ||
| Overweight | 0.86 (0.65–1.13) | 0.90 (0.69–1.16) |
| Obese | 1.59 (1.27–1.99)*** | 0.98 (0.78–1.22) |
OR = Odds Ratio. CI = Confidence Interval. Ref = Reference. HEI = Health Eating Index. MET = Metabolic Equivalent of Task.
P<0.10.
P <0.05.
P<0.01.
P<0.001.
The NHANES complex sampling design was accounted for the logistic regression. The c-statistic of the multivariable model was 0.831 (95% CI=0.818–0.842). The variance inflation factors of all variables in the models were lower than 10, indicating no multicollinearity between variables.
For the other control variables, sex, age, race/ethnicity, marital status, income, health insurance, chronic condition, general health, and smoking were significantly associated with depression. Men were less likely to be depressed than women (OR=0.63). Elders were less likely to be depressed than adults aged 20–44 years (OR=0.55). People who were never married and in other status were more likely to be depressed than people who are married or cohabitating (ORs=1.68 and 1.45, respectively). People of high income were less likely to be depressed than people in poverty (OR=0.56). People with private insurance were less likely to be depressed than people without insurance (OR=0.67). Compared with people who reported excellent or very good general health, people who reported good general health had a higher depression rate (OR=2.27). People who reported fair or poor general health had the highest depression rate than those who reported excellent or very good general health (OR=5.43). Current smokers were also more likely to be depressed than people who never smoked (OR=1.55). People with history of sleep disorder were more likely to be depressed than their counterparts (OR=3.20).
We investigated bivariate associations between race/ethnicity and the main and controlling variables (see Appendix 2). All variables were significantly associated with race/ethnicity. To investigate the effect of these variables by race/ethnicity, we conducted multivariable logistic regression as subgroup analyses (Table 3). Regarding diet or physical activity variables, significant association was found in diet-related variables among only Whites and Blacks. Whites who received an acceptable HEI total were less likely to be depressed than people who received a poor HEI total (OR=0.11). Blacks who reported they have poor/fair dietary pattern were more likely to be depressed than those who reported excellent/very good self-rated dietary pattern (OR=1.88). Regarding the other control variables, history of sleep disorder and general health status were associated with depression for all racial/ethnic groups. People with history of sleep disorder had higher rates of depression than people without it, ranging from the lowest OR of 2.51 for Mexican Hispanic to the highest OR of 4.44 for Asians. People who reported fair or poor general health had higher depression rates than people who reported excellent or very good general health, with ORs ranging from 3.41 (for Black) to 8.82 (for Other Hispanic). Specific to Whites, sex, marital status, and income were significantly associated with depression. Whites who were never married were more likely to be depressed than people who were married or cohabitating (OR=2.21). Specific to Asians, sex, age, and smoking were associated with depression. Asians aged 65 years or older were less likely to be depressed than younger people aged 20–44 years old (OR=0.06). Asian previous smokers were more likely to be depressed than people who never smoked (OR=5.07) and current smokers (OR=3.54). Specific to Blacks, sex, age, marital status, income, obesity, and smoking were associated with depression. Of note, obese Blacks were less likely to be depressed than people who have normal weight or are underweight (OR=0.60). Specific to Mexican Hispanics, marital status, health insurance, and number of chronic conditions were associated with depression. Mexican Hispanics with private insurance were less likely to be depressed than people without insurance (OR=0.31), and people with multiple chronic conditions were more likely to be depressed than people with no chronic conditions (OR=4.99). Among Other Hispanics, sex and smoking were associated with depression. Women were less likely to be depressed than men (OR=0.42) and current smokers were more likely to be depressed than people who never smoked (OR=2.24).
Table 3.
Odds Ratio (and 95% Confidence Interval) of Logistic Regression on Depression by Race/Ethnicity
| Variable | White | Asian | Black | Mexican | Other Hispanic |
|---|---|---|---|---|---|
| Main Variable | |||||
| Self-rated dietary pattern (Ref: Excellent/very good) | |||||
| Good | 1.16 (0.75–1.82) | 1.36 (0.60–3.08) | 1.08 (0.70–1.67) | 0.38 (0.14–1.04) | 1.15 (0.49–2.70) |
| Fair/poor | 1.62 (0.93–2.82) | 0.89 (0.28–2.80) | 1.88 (1.18–2.99)* | 0.80 (0.36–1.78) | 1.69 (0.72–4.00) |
| HEI Total (Ref: Poor) | |||||
| Need improvement | 0.84 (0.61–1.17) | 1.14 (0.42–3.10) | 1.12 (0.78–1.62) | 0.83 (0.51–1.36) | 1.31 (0.75–2.26) |
| Acceptable | 0.11 (0.02–0.75)* | 1.84 (0.58–5.84) | 0.72 (0.16–3.24) | 0.34 (0.09–1.34) | 0.92 (0.23–3.69) |
| Caffeine (Ref: None) | |||||
| 1–200 mg | 0.79 (0.46–1.36) | 1.05 (0.30–3.60) | 1.20 (0.79–1.83) | 0.42 (0.24–1.36)+ | 1.68 (0.80–3.50) |
| 201–400 mg | 0.80 (0.43–1.50) | 1.52 (0.30–7.80) | 1.63 (0.86–3.12) | 0.32 (0.09–1.34)+ | 1.75 (0.67–4.60) |
| >400 mg | 0.84 (0.44–1.60) | 4.79 (0.80–28.60) | 1.00 (0.44–2.26) | 0.89 (0.24–3.23) | 1.23 (0.47–3.19) |
| Total Leisure-Time Physical Activity (Ref: None) | |||||
| Low | 0.93 (0.64–1.37) | 1.21 (0.50–2.92) | 0.93 (0.55–1.57) | 0.96 (0.49–1.86) | 0.62 (0.29–1.35) |
| Moderate | 0.83 (0.53–1.29) | 0.80 (0.25–2.51) | 0.94 (0.48–1.84) | 0.74 (0.29–1.87) | 0.48 (0.22–1.07) |
| High | 0.57 (0.32–1.02)+ | 1.74 (0.48–6.28) | 0.65 (0.40–1.04) | 0.61 (0.19–1.90) | 1.09 (0.41–2.85) |
| Socio-Demographic | |||||
| Sex (Ref: Female) | |||||
| Male | 0.60 (0.46–0.78)** | 0.47 (0.23–0.98)* | 0.68 (0.48–0.97)* | 0.67 (0.38–1.18) | 0.42 (0.23–0.76)* |
| Age (Ref: 20–44 yrs.) | |||||
| 45–64 yrs. | 1.12 (0.78–1.60) | 0.70 (0.32–1.51) | 0.46 (0.31–0.68)** | 0.78 (0.42–1.46) | 0.95 (0.52–1.73) |
| ≥65 yrs. | 0.60 (0.36–1.01)+ | 0.06 (0.01–0.38)* | 0.35 (0.19–0.66)** | 0.72 (0.28–1.83) | 0.40 (0.12–1.31) |
| Education (Ref: <High school) | |||||
| High school | 0.82 (0.52–1.29) | 1.09 (0.23–5.01) | 0.75 (0.52–1.09) | 1.18 (0.57–2.41) | 0.87 (0.47–1.63) |
| Some college | 0.89 (0.56–1.41) | 0.84 (0.18–3.84) | 0.86 (0.51–1.45) | 1.13 (0.66–1.92) | 1.08 (0.60–1.94) |
| College | 0.72 (0.38–1.37) | 1.13 (0.31–4.17) | 0.78 (0.42–1.46) | 1.68 (0.42–6.79) | 0.79 (0.38–1.66) |
| Marital Status (Ref: Married/cohabitating) | |||||
| Never married | 2.21 (1.47–3.33)** | 1.15 (0.44–2.98) | 1.41 (0.98–2.02)+ | 1.09 (0.53–2.23) | 1.09 (0.62–1.92) |
| Other | 1.43 (1.06–1.93)* | 2.20 (0.95–5.09)+ | 2.17 (1.51–3.14)** | 1.75 (1.06–2.87)* | 0.98 (0.60–1.62) |
| Income (Ref: Poor) | |||||
| Low | 0.94 (0.66–1.33) | 1.11 (0.27–4.54) | 0.85 (0.59–1.24) | 0.72 (0.46–1.11) | 0.70 (0.42–1.17) |
| Moderate | 0.87 (0.56–1.33) | 2.28 (0.84–6.20) | 0.60 (0.39–0.92)* | 0.77 (0.48–1.23) | 0.80 (0.38–1.70) |
| High | 0.52 (0.30–0.91)* | 2.79 (0.99–7.87)+ | 0.33 (0.14–0.76)* | 0.83 (0.19–3.68) | 0.78 (0.19–3.17) |
| Missing | 0.65 (0.31–1.36) | 0.98 (0.22–4.54) | 0.69 (0.34–1.40) | 1.02 (0.55–1.88) | 0.69 (0.29–1.65) |
| Health Insurance (Ref: No insurance) | |||||
| Public | 1.46 (0.93–2.29) | 1.93 (0.46–8.05) | 0.74 (0.50–1.10) | 0.91 (0.46–1.82) | 1.43 (0.70–2.92) |
| Private | 0.75 (0.53–1.06) | 0.71 (0.24–2.05) | 0.71 (0.44–1.15) | 0.31 (0.13–0.74)* | 0.69 (0.28–1.74) |
| Other | 1.56 (0.91–2.68) | 0.47 (0.10–2.17) | 0.89 (0.50–1.61) | 1.94 (0.63–5.93) | 0.48 (0.22–1.04)+ |
| Health-related | |||||
| # of Chronic Conditions (Ref: 0) | |||||
| 1 | 0.90 (0.62–1.30) | 1.55 (0.55–4.37) | 1.12 (0.75–1.69) | 1.12 (0.62–2.03) | 0.82 (0.49–1.37) |
| 2+ | 1.18 (0.78–1.77) | 2.98 (0.96–9.23)+ | 1.42 (0.83–2.42) | 4.99 (2.03–12.26)* | 1.17 (0.62–2.20) |
| General Health (Ref: Excellent/very good) | |||||
| Good | 2.44 (1.57–3.77)** | 1.30 (0.57–3.00) | 1.24 (0.86–1.78) | 1.56 (0.53–4.64) | 3.20 (1.49–6.87)* |
| Fair/poor | 5.94 (3.81–9.26)*** | 7.77 (3.54–17.08)*** | 3.41 (2.11–5.50)*** | 3.50 (1.18–10.37)* | 8.82 (3.93–19.78)*** |
| Behavioral-related | |||||
| Alcohol | 1.04 (0.76–1.41) | 0.53 (0.20–1.39) | 1.21 (0.87–1.69) | 1.29 (0.83–2.00) | 1.07 (0.60–1.94) |
| Smoking (Ref: Never) | |||||
| Previous | 0.88 (0.57–1.36) | 5.07 (1.60–16.04)* | 1.47 (0.89–2.43) | 1.17 (0.61–2.23) | 0.89 (0.44–1.80) |
| Current | 1.32 (0.90–1.92) | 3.54 (1.06–11.77)* | 1.78 (1.21–2.62)* | 1.83 (0.89–3.75) | 2.24 (1.11–4.51)* |
| Sleep disorder | 3.50 (2.63–4.68)*** | 4.44 (1.83–10.76)** | 2.86 (2.20–3.71)*** | 2.51 (1.80–3.49)* | 2.65 (1.86–3.78)*** |
| Obesity (Ref: Normal/underweight) | |||||
| Overweight | 0.92 (0.65–1.29) | 1.42 (0.85–2.37) | 0.88 (0.61–1.25) | 0.59 (0.22–1.61) | 1.10 (0.51–2.34) |
| Obese | 0.95 (0.70–1.30) | 0.40 (0.11–1.49) | 0.60 (0.43–0.82)* | 1.04 (0.54–2.00) | 1.25 (0.65–2.41) |
HEI = Health Eating Index. MET = Metabolic Equivalent of Task. Ref = Reference.
P<0.10.
P <0.05.
P<0.01.
P<0.001.
The NHANES complex sampling design was considered for all the analyses. The variance inflation factors of all variables in the models were lower than 10, indicating no multicollinearity between variables.
The forest plot in Figure 1 depicts ORs and 95% CI for race/ethnicity and the significant variables in the multivariable logistic regression on use of mental health specialists among depressed people. Although all racial/ethnic groups showed lower rates in use of mental health specialists among depressed people compared to Whites, race/ethnicity was not significant. The significant variables associated with use of mental health specialists included caffeine intake, age, education, health insurance, chronic condition, and history of sleep disorder. Older people were less likely to visit mental health specialists compared to people aged 20–44 years (45–64 years: OR=0.52; 65 years or older: OR=0.22). People who drank 0–200 mg of coffee per day (about 1 cup of coffee) were 1.79 times more likely to visit mental health specialists than people who do not drink at all. People having at least a college degree and high school graduates were 3.10 and 1.85 times more likely to visit mental health specialists than people having less than a high school education, respectively. People with public insurance were 3.28 times more likely to visit mental health specialists than people without insurance. People with two or more chronic conditions and history of sleep disorder were 1.98 and 3.09 times more likely to use mental health specialists than their counterparts, respectively.
Figure 1.
Factors Associated with Use of Mental Health Professionals among Depressed People. Multivariable logistic regression was conducted on use of mental health professionals among depressed people adjusting for the NHANES complex sampling design. The variance inflation factors of all variables in the model were lower than 10, indicating no multicollinearity between variables. CI = confidence interval.
Discussion
In this study, we investigated the effect of race/ethnicity, dietary intake, and physical activity on depression adjusting for other potential factors and identified important variables associated with the use of mental health specialists among depressed people using the NHANES 2011–2016 data. We found that dietary intake assessed by self-rated dietary pattern and HEI total showed significant associations with depression. This result implies poor dietary intake may be one of the risk factors for depression and dietary factors can be targeted in efforts to decrease depression. Other studies also showed that healthy eating patterns was negatively associated with depression [30]. A recent randomized controlled trial of an adjunctive dietary intervention in the treatment of moderate-to-severe depression proved that dietary improvement programs can reduce major depressive episodes [31]. Therefore, clinicians or mental health specialists should consider the dietary patterns of their patients with depression, and advocate for improvement in their patients’ dietary patterns.
One may ask, “How can we improve diet to reduce or prevent depression?” Several studies have explained dietary patterns that may reduce risk of depression by investigating certain dietary patterns or food types. For example, the Mediterranean-style diet, rich in fruit, vegetables, nuts, and cereals, has been reported to lower the risk of depression or is negatively associated with depression [32, 33]. The mechanism between depression and Mediterranean diet has not been well understood but it may be explained by the relationship of reduced interleukin (IL)-6 level with adherence to the Mediterranean diet [34]. This mechanism is also supported by a recent longitudinal study showing a negative association between IL-6 level and depression [35]. Besides Mediterranean diet, other dietary patterns or food types were also found to be associated with depression. For instance, there is evidence that people with depression consume more fat and sugar [36] but less fruits and vegetables [37] or walnuts [38] than people without depression. Our data also suggest that depressed people consumed higher sugar, fewer amounts of fruits and vegetables and lower fatty acid (see Appendix 1). According to a recent study, gut microbiota can influence food cravings and eating behaviors [39]. Hence, we speculate that gut microbiota affected by diet can affect stress reaction and lead to depression. Future studies are needed to address this potential modulation between gut microbiota and depression by an intervention study.
Regarding physical activity, we found a significant association between physical activity and depression in the multivariable model. De Mello et al. (2013) also showed that the odds of being depressed among people who do not practice exercise was 40% higher than among those people who do at least three times of physical activity per week [8]. Systematic review and meta-analysis studies concluded physical activity has a protective effect on depression [10, 40, 41]. These studies with our finding infer moderate physical activity can serve as a valuable strategy in reducing depression level or preventing depression.
Our study found the prevalence of depression was 8.2% among the US adult population and the rates vary by race/ethnicity ranging from 3.1% (Asian) to 15.0% (Other). Even after controlling for the other variables, Asians, Blacks and Mexican Hispanics were less likely to have depression than Whites. This is consistent with the findings from other studies [11–15]. Our data are representative to the US population and the PHQ-9 is a validated tool to define depression, tested measurement invariance regarding ethnicity for various populations [42–44]. Therefore, the lower prevalence rates of depression in these racial/ethnic subpopulations may reflect the reality. However, we should be cautious of interpreting this result because cultural-based expressions may affect the differences in depression symptoms [45]. There may be a need to develop a tool to diagnose depression incorporating cultural aspects.
Our subgroup analyses showed that self-rated general health and history of sleep disorder were significantly associated with depression across all the racial/ethnic groups. Previous literatures recognized a strong and independent association between depression and self-rated general health [46] and between depression and sleep disorders [47]. Thus, we suggest that clinicians should check whether their patients report fair/poor self-rated health or suffer a sleep disorder. The clinicians can advise the patients with one or both conditions to take a depression diagnosis test and refer them to a mental health specialist if depression is suspected.
The subgroup analyses also identified different variables associated with depression by race/ethnicity. Dietary intake was associated with depression among Whites and Blacks only. This modifiable factor can be targeted for tailored interventions specific to these racial groups to reduce current depression or prevent depression. Clinicians and mental health specialists should suggest their White and Black patients with depression to improve their dietary patterns.
Asian group delineated different patterns for age and smoking status compared to the other racial/ethnic groups. For age, the odds of being depressed among Asian elders aged 65 years or older were the smallest among all the racial/ethnic groups compared to the adults aged 20–44 years. We suspect that this may be related to Asian cultures such as Confucianism. Confucianism is a system that holds a set of moral behaviors designed to regulate the relationships of societies and emphasize familial and social harmony [48]. Hence, the older Asian participants raised in Confucian-based societies might have hid their genuine negative feelings but selected more modest responses. A study is needed to explain the link between aging and depression for Asians.
Mexican and Other Hispanics showed different patterns in variables associated with depression. For Mexican Hispanics, marital status, health insurance, and chronic condition were significantly associated with depression, while sex and smoking were significantly associated for Other Hispanics. The lower OR of depression among Mexican Hispanics with private insurance shows that they might have a higher income or have stable jobs. In literature, Mexican Hispanics are frequently combined with Other Hispanics, but our study showed they may have different association patterns with a disease. Hispanics have various origins and histories. People from Latino countries with a history of political violence could have experienced multiple traumas [49], which may affect the individuals’ mental or physical health. Therefore, more research is needed to identify its contributing factors for depression among Hispanics based on their country origins and history of political violence.
Racial/ethnic disparities in mental health service utilization have been widely investigated. Studies have found that racial/ethnic minority groups reported lower utilization rates than Whites [19, 21–23]. Regarding mental health services utilization among depressed people, we found only 26% of depressed people visited mental health specialists. Although the use of mental health specialists was lower for racial/ethnic minority groups than Whites among depressed people in our study, there were no significant differences in mental health services utilization among racial/ethnic groups. In an extra analysis, we found that there is a significant racial/ethnic difference in using mental health specialists among non-depressed people. This alludes that there are no racial/ethnic disparities in mental health services utilization among people with depression. The racial/ethnic differences in mental health utilization might be from people with other types of mental disorders.
Socioeconomic factors such as education and health insurance appeared to be more important barriers to mental health services utilization. Remarkably, depressed people with public insurance (e.g., Medicare, Medicaid) were more likely to visit mental health specialists. The coverage for mental health services by public insurance plans (e.g., Medicaid Alternative Benefit Plan) seemed to help patients with depression use more of this service. Since mental health services are now considered essential services that are covered by health insurance [50], future studies should be conducted to investigate whether unmet needs for mental health utilization among patients with depression has been improved since the Affordable Care Act was enacted.
Several limitations should be considered when interpreting the results. First, the data are cross-sectional. Causal effects between variables could not be tested in this study. The focus of this study was to identify risk or protective factors of depression including nutrition and exercise. Second, the measure for the use of mental health specialists may not reflect the actual utilization. The PHQ-9 to assess depression was measured at the time of the MEC visit, but the question for use of mental health service asks about the past 30 days. Third, missing data may have affected the analysis results, although the overall missing rate was relatively small (<5%). Fourth, important risk factors may not have been included in the analysis. We included many candidates but there still might be other important variables for depression such as a family history of depression and substance misuse. There is a questionnaire for substance use included in NHANES, but it is limited to adults aged between 18 and 59 years. Including the substance use questions and restricting the age group of the adults could generate different results from this study. Fifth, the 24-hour food diary recall is self-reported, which can be subject to response bias and measurement errors. Sixth, the use of HEI total score can be another limitation. Although the HEI is the most frequently used tool to measure individuals’ overall diet quality independent of quantity, aligning with the 2015–2020 Dietary Guidelines for Americans [26], the HEI score may not be optimal for all racial and ethnic groups because it does not reflect various cultures and availabilities in food. We found some differences in all HEI components among racial/ethnic groups (see Appendix 3). Future study is needed to investigate how culture affects diet quality and choices. In addition, we found a high correlation between two fruit HEI components (r = 0.82), indicating multicollinearity. Although the HEI has been validated [26, 51–52], further study may be necessary to explore other possibilities of different or equivalent clustering of diet factors. Seventh, the categories of race/ethnicity were based on the NHANES available categorization, where diverse Asian ethnic groups were merged into a single Asian group. Finally, there may be other important variables that are not included in this study or hidden clustering of health or behavioral variables. Even though investigating variables other than physical activity and diet is not the main focus of this paper, we included other health or behavioral variables such as alcohol consumption and history of sleep disorder. These variables were treated as individual items rather than constructing any latent clusters in our analyses, because they were not highly correlated (r < 0.30).
Despite these limitations, our study has several strengths. First, the use of a representative sample dataset makes it possible to generalize our findings to the general US adult population. Second, looking at the many risk factors simultaneously, our study provides a more comprehensive view about depression. Third, depression was clinically assessed by a validated and reliable nine-item instrument, the PHQ-9. Lastly, the use of disaggregated Asian and Mexican Hispanic ethnicities, considered as Other or Hispanic ethnic groups most times, allows us to identify differences in depression.
Conclusions
Prevalence of depression was lower among those who have good or acceptable diet and do moderate physical activity. Clinicians could consider recommending improvement of the habitual diet and physical activity to their patients with depression. Future psychotherapeutic intervention research for depression should consider incorporating these modifiable factors.
Acknowledgements:
This study was partially supported by U54MD007601 (Ola HAWAII) from the National Institutes of Minority Health and Health Disparities (NIMHD). We appreciate the National Health and Nutrition Examination Survey (NHANES) participants who took the time to share their information on socio-demographic, health and eating behavior; and the efforts of the Centers for Disease Control and Prevention (CDC) to collect data and make this readily accessible. We also thank Ms. Munirih Taafaki for editing this manuscript.
Funding. U54MD007601 (Ola HAWAII)
Appendix 1.
Bivariate Association of Nutrition-Related Variables with Depression
| Variable | Total Weighted Mean (SE) | Depression Status | P-value | |
|---|---|---|---|---|
| Not depressed Weighted Mean (SE) | Depressed Weighted Mean (SE) | |||
| 24hr-Recall Diet | ||||
| Calorie (kcal) | 2145.1 (10.5) | 2159.0 (11.7) | 2024.8 (43.5) | 0.005 |
| Protein (gm) | 83.0 (0.5) | 84.2 (0.6) | 71.0 (1.9) | <0.001 |
| Carbohydrate (mg) | 254.1 (1.3) | 254.5 (1.4) | 250.7 (5.2) | 0.493 |
| Sugar (mg) | 111.3 (0.8) | 110.4 (0.9) | 121.1 (4.0) | 0.014 |
| Fiber (mg) | 17.5 (0.2) | 17.8 (0.2) | 14.7 (0.4) | <0.001 |
| Fat (mg) | 83.1 (0.6) | 84.0 (0.6) | 75.6 (1.9) | <0.001 |
| Cholesterol (mg) | 291.9 (2.9) | 294.8 (3.1) | 251.5 (9.2) | <0.001 |
| α-tocopherol (mg) | 9.2 (0.1) | 9.4 (0.1) | 8.1 (0.4) | <0.001 |
| Retinol (mcg) | 431.6 (9.2) | 436.0 (10.3) | 403.0 (18.6) | 0.115 |
| Vitamin A (mcg) | 646.5 (13.9) | 659.6 (15.3) | 534.7 (19.9) | <0.001 |
| α-carotene (mcg) | 422.8 (21.4) | 441.0 (23.1) | 243.3 (24.9) | <0.001 |
| β-carotene (mcg) | 2,332.6 (77.0) | 2,427.0 (81.4) | 1,434.1 (87.7) | <0.001 |
| β-cryptoxanthin (mcg) | 84.8 (2.5) | 86.1 (2.7) | 60.7 (4.4) | <0.001 |
| Vitamin B1 (mg) | 1.63 (0.01) | 1.63 (0.01) | 1.43 (0.01) | <0.001 |
| Iron (mg) | 14.8 (0.1) | 14.9 (0.1) | 13.3 (0.4) | <0.001 |
| Phosphorus (mg) | 1,400.5 (9.2) | 1,416 (9.6) | 1,246 (32.4) | <0.001 |
| Caffeine (mg) | 169.8 (4.2) | 168.4 (4.4) | 205.4 (14.8) | 0.014 |
| Healthy Eating Index | ||||
| Total vegetable | 3.26 (0.02) | 3.30 (0.02) | 2.90 (0.06) | <0.001 |
| Green/Bean | 2.03 (0.03) | 2.07 (0.03) | 1.62 (0.08) | <0.001 |
| Fruit | 2.35 (0.03) | 2.37 (0.04) | 1.95 (0.08) | <0.001 |
| Whole Fruit | 2.52 (0.04) | 2.55 (0.04) | 2.02 (0.10) | <0.001 |
| Whole Grain | 2.96 (0.05) | 3.02 (0.05) | 2.34 (0.11) | <0.001 |
| Dairy | 5.33 (0.06) | 5.33 (0.06) | 5.32 (0.17) | 0.936 |
| Protein | 4.48 (0.01) | 4.50 (0.01) | 4.26 (0.05) | <0.001 |
| Sea Plant | 2.89 (0.03) | 2.94 (0.03) | 2.49 (0.09) | <0.001 |
| Fatty Acid | 5.04 (0.05) | 5.07 (0.05) | 4.68 (0.17) | 0.036 |
| Sodium | 3.97 (0.04) | 3.91 (0.05) | 4.55 (0.13) | <0.001 |
| Refined Grain | 6.35 (0.06) | 6.37 (0.06) | 6.32 (0.16) | 0.786 |
| Saturated Fat | 5.96 (0.06) | 5.95 (0.06) | 5.91 (0.14) | 0.779 |
| Added Sugar | 6.94 (0.06) | 7.01 (0.05) | 6.07 (0.20) | <0.001 |
| Total Score | 54.08 (0.27) | 54.39 (0.28) | 50.43 (0.51) | <0.001 |
SE = Standard Error.
Two sample t test was used to explore bivariate association with depression, adjusting for the NHANES complex sampling design.
Health Eating Index: The possible ranges of all components are 0–10 except total vegetable, green/bean, fruit, whole fruit, sea plant, and protein whose range is 0–5. The total score ranges 0–100. A higher score of the Health Eating Index component indicates a better quality of the dietary component.
Appendix 2.
Bivariate Association with Race/Ethnicity
| Variable | Race/Ethnicity, Weighted column % | P-value | |||||
|---|---|---|---|---|---|---|---|
| White | Asian | Black | Mexican | Other Hispanic | Other | ||
| Main Variable | |||||||
| Self-rated dietary pattern | <0.001 | ||||||
| Excellent/very good | 34.9 | 43.0 | 24.6 | 13.7 | 21.7 | 27.6 | |
| Good | 42.8 | 41.8 | 40.4 | 40.2 | 44.2 | 42.4 | |
| Fair/poor | 22.3 | 15.1 | 35.0 | 46.1 | 34.2 | 30.0 | |
| HEI Total | <0.001 | ||||||
| Poor | 66.0 | 50.3 | 74.8 | 72.8 | 63.1 | 72.9 | |
| Need improvement | 30.8 | 42.8 | 23.1 | 24.6 | 32.7 | 26.0 | |
| Acceptable | 3.2 | 6.8 | 2.1 | 2.6 | 4.1 | 1.0 | |
| Caffeine | <0.001 | ||||||
| None | 9.1 | 17.0 | 28.8 | 17.0 | 13.3 | 9.7 | |
| 1–200 mg | 51.9 | 67.6 | 60.3 | 64.3 | 65.3 | 58.2 | |
| >200–400 mg | 26.3 | 13.0 | 8.8 | 14.6 | 16.5 | 19.3 | |
| >400 mg | 12.7 | 2.5 | 2.1 | 4.1 | 4.9 | 12.7 | |
| Total Leisure-Time Physical Activity (MET-hours/week) | <0.001 | ||||||
| None | 43.2 | 39.2 | 49.3 | 53.7 | 50.3 | 44.4 | |
| Low (≤12) | 21.6 | 23.5 | 19.3 | 17.1 | 18.4 | 20.5 | |
| Moderate (>12 to 32) | 21.2 | 22.9 | 14.8 | 15.6 | 17.5 | 14.9 | |
| High (>32) | 13.9 | 14.3 | 16.7 | 13.6 | 13.8 | 20.1 | |
| Socio-Demographic | |||||||
| Sex | 0.016 | ||||||
| Female | 51.2 | 53.2 | 55.0 | 49.1 | 53.2 | 47.4 | |
| Male | 48.8 | 46.8 | 45.0 | 50.9 | 46.8 | 52.6 | |
| Age | <0.001 | ||||||
| 20–44 | 38.7 | 53.0 | 50.7 | 63.7 | 57.4 | 54.8 | |
| 45–64 | 38.4 | 32.4 | 36.3 | 28.5 | 31.3 | 30.3 | |
| ≥65 | 22.9 | 14.6 | 13.0 | 7.8 | 11.3 | 15.0 | |
| Education | <0.001 | ||||||
| <High school | 9.53 | 11.0 | 18.4 | 45.1 | 29.6 | 10.3 | |
| High school | 20.9 | 13.3 | 25.9 | 20.7 | 22.1 | 20.4 | |
| Some college | 34.1 | 20.6 | 36.9 | 25.0 | 30.8 | 43.8 | |
| College | 35.5 | 55.0 | 18.8 | 9.2 | 17.5 | 25.6 | |
| Marital Status | <0.001 | ||||||
| Married/cohabitating | 64.7 | 67.2 | 41.2 | 67.4 | 57.9 | 57.8 | |
| Never married | 16.1 | 22.8 | 35.8 | 19.1 | 22.6 | 21.2 | |
| Other (divorced. separated, widowed) | 19.2 | 9.9 | 22.9 | 13.5 | 19.4 | 21.0 | |
| Ratio of Family Income to Poverty | <0.001 | ||||||
| Poor | 10.2 | 13.9 | 26.2 | 27.2 | 26.8 | 18.2 | |
| Low | 17.8 | 15.0 | 23.6 | 29.2 | 23.7 | 24.4 | |
| Moderate | 26.8 | 25.2 | 23.7 | 22.6 | 25.4 | 26.0 | |
| High | 40.4 | 37.2 | 17.0 | 9.8 | 14.5 | 23.3 | |
| Missing | 4.9 | 8.6 | 9.6 | 11.2 | 9.6 | 8.1 | |
| Health Insurance | <0.001 | ||||||
| No insurance | 11.3 | 15.8 | 21.6 | 44.1 | 31.1 | 21.2 | |
| Public | 13.3 | 13.0 | 24.3 | 14.2 | 19.6 | 15.7 | |
| Private | 69.0 | 63.0 | 44.6 | 34.0 | 41.5 | 50.3 | |
| Other | 6.4 | 8.2 | 9.5 | 7.6 | 7.8 | 12.7 | |
| Health-related | |||||||
| # of Chronic Conditions | <0.001 | ||||||
| 0 | 43.6 | 65.6 | 52.0 | 63.1 | 56.9 | 41.2 | |
| 1 | 27.5 | 22.7 | 25.9 | 22.9 | 26.7 | 29.2 | |
| 2+ | 28.9 | 11.7 | 22.0 | 13.9 | 16.4 | 29.5 | |
| General Health | <0.001 | ||||||
| Excellent/very good | 52.0 | 49.3 | 37.6 | 224.3 | 34.4 | 41.9 | |
| Good | 33.6 | 38.2 | 39.9 | 39.9 | 39.1 | 35.6 | |
| Fair/poor | 14.4 | 12.4 | 22.5 | 35.7 | 26.5 | 22.5 | |
| Behavioral-related | |||||||
| Alcohol | <0.001 | ||||||
| No | 63.1 | 81.1 | 67.2 | 52.9 | 57.5 | 63.4 | |
| Yes | 36.9 | 18.9 | 32.8 | 47.1 | 42.5 | 36.6 | |
| Smoking | <0.001 | ||||||
| Never | 52.2 | 75.7 | 59.6 | 66.1 | 62.0 | 45.3 | |
| Previous | 28.5 | 14.2 | 15.3 | 18.9 | 21.4 | 22.3 | |
| Current | 19.3 | 10.1 | 25.1 | 15.0 | 16.6 | 32.4 | |
| Sleep disorder | <0.001 | ||||||
| No | 67.4 | 84.9 | 74.4 | 81.6 | 77.8 | 65.3 | |
| Yes | 32.6 | 15.1 | 25.6 | 18.4 | 22.2 | 34.7 | |
| Obesity | <0.001 | ||||||
| Normal or underweight | 29.9 | 57.6 | 24.8 | 19.0 | 25.1 | 29.2 | |
| Overweight | 34.2 | 29.7 | 26.8 | 35.0 | 34.4 | 27.1 | |
| Obese | 35.9 | 12.7 | 48.4 | 46.0 | 40.5 | 43.8 | |
HEI = Health Eating Index. MET = Metabolic Equivalent of Task. Adjusting for the NHANES complex sampling design, Rao-Scott’s chi-square test was used to explore bivariate association with race/ethnicity.
Appendix 3.
Bivariate Association between Nutrition-Related Variables and Race/Ethnicity
| Variable | Race/Ethnicity, Weighted Mean (SE) | |||||
|---|---|---|---|---|---|---|
| White | Asian | Black | Mexican | Other Hispanic | Other | |
| 24hr-Recall Diet | ||||||
| Calorie (kcal) | 2148.6 (13.7) | 1900.9 (21.9)* | 2156.5 (26.9) | 2265.4 (30.1)* | 2098.8 (28.8) | 2215.5 (57.7) |
| Protein (gm) | 82.3 (0.7) | 81.7 (1.1) | 80.3 (0.9) | 91.2 (1.5)* | 85.1 (1.5) | 84.5 (2.6) |
| Carbohydrate (mg) | 251.1 (2.1) | 238.0 (2.8)* | 257.7 (3.3) | 273.7 (4.2)* | 259.1 (3.5)* | 268.4 (8.1)* |
| Sugar (mg) | 111.9 (1.3) | 83.0 (1.6)* | 118.0 (2.2)* | 113.2 (2.3) | 111.9 (2.1) | 118.9 (5.1) |
| Fiber (mg) | 17.4 (0.3) | 19.1 (0.4)* | 14.9 (0.3)* | 20.7 (0.5)* | 17.5 (0.4) | 17.8 (0.9) |
| Fat (mg) | 84.3 (0.7) | 67.3 (1.2)* | 83.2 (1.2) | 86.2 (1.2) | 77.2 (1.6)* | 84.5 (2.6) |
| Cholesterol (mg) | 282.4 (3.3) | 264.7 (5.9)* | 309.5 (5.0)* | 348.0 (7.3)* | 304.7 (8.3)* | 292.3 (11.9) |
| α-tocopherol (mg) | 9.5 (0.2) | 8.5 (0.2)* | 8.8 (0.2)* | 8.6 (0.2)* | 8.1 (0.2)* | 9.8 (0.6) |
| Retinol (mcg) | 465.8 (12.2) | 277.9 (8.6)* | 351.6 (7.0)* | 412.5 (13.8)* | 376.4 (10.7)* | 433.1 (34.8) |
| Vitamin A (mcg) | 684.6 (18.9) | 631.7 (17.8)* | 551.5 (13.7)* | 575.8 (17.5)* | 548.0 (16.5)* | 602.1 (38.7)* |
| α-carotene (mcg) | 446.6 (31.2) | 661.4 (29.9)* | 282.6 (29.1)* | 318.4 (20.4)* | 391.1 (38.8) | 362.6 (70.7) |
| β-carotene (mcg) | 2371.5 (105.5) | 3839.6 (168.6)* | 2225.3 (128.0) | 1759.1 (99.3)* | 1820.8 (121.2)* | 1818.8 (214.1)* |
| β-cryptoxanthin (mcg) | 76.3 (2.7) | 163.7 (14.7)* | 74.5 (3.3) | 107.0 (7.0)* | 100.7 (8.9)* | 68.3 (6.5) |
| Vitamin B1 (mg) | 1.64 (0.01) | 1.59 (0.02) | 1.47 (0.02)* | 1.69 (0.03) | 1.56 (0.03)* | 1.59 (0.05) |
| Iron (mg) | 14.9 (0.1) | 14.1 (0.3)* | 13.6 (0.2)* | 16.3 (0.4)* | 14.1 (0.3)* | 14.6 (0.5) |
| Phosphorus (mg) | 1418.4 (10.9) | 1260.7 (17.3)* | 1272.2 (14.6)* | 1542.3 (26.4)* | 1369.6 (24.1)* | 1400.9 (42.1) |
| Caffeine (mg) | 202.2 (4.8) | 100.0 (4.0)* | 76.8 (3.3)* | 113.5 (4.4)* | 126.8 (6.9)* | 187.6 (14.5) |
| Healthy Eating Index | ||||||
| Total vegetable | 3.28 (0.03) | 3.72 (0.05)* | 2.94 (0.04)* | 3.30 (0.05) | 3.27 (0.04) | 3.04 (0.11)* |
| Green/Bean | 1.94 (0.05) | 2.91 (0.06)* | 1.70 (0.05)* | 2.42 (0.07)* | 2.44 (0.06)* | 1.77 (0.02) |
| Fruit | 2.29 (0.05) | 2.99 (0.06)* | 2.22 (0.06) | 2.46 (0.06)* | 2.68 (0.06)* | 1.99 (0.12)* |
| Whole Fruit | 2.55 (0.05) | 3.29 (0.06)* | 1.95 (0.07)* | 2.60 (0..08) | 2.61 (0.08) | 2.15 (0.16)* |
| Whole Grain | 3.15 (0.05) | 3.73 (0.14)* | 2.44 (0.09)* | 2.13 (0.08)* | 2.41 (0.13)* | 2.90 (0.20) |
| Dairy | 5.72 (0.05) | 4.10 (0.12)* | 3.94 (0.07)* | 5.25 (0.09)* | 5.14 (0.11)* | 4.90 (0.19)* |
| Protein | 4.44 (0.02) | 4.56 (0.03)* | 4.56 (0.02)* | 4.54 (0.03)* | 4.59 (0.03)* | 4.47 (0.05) |
| Sea Plant | 2.88 (0.04) | 3.67 (0.06)* | 2.52 (0.07)* | 2.97 (0.08) | 3.01 (0.06) | 2.76 (0.17) |
| Fatty Acid | 4.79 (0.06) | 6.42 (0.10)* | 5.74 (0.09)* | 4.92 (0.11) | 5.26 (0.10)* | 5.20 (0.24) |
| Sodium | 4.01 (0.06) | 2.69 (0.12)* | 4.17 (0.07) | 3.99 (0.12) | 4.29 (0.09)* | 4.10 (0.24) |
| Refined Grain | 6.69 (0.07) | 5.32 (0.12) | 6.78 (0.08) | 4.30 (0.14) | 5.73 (0.16) | 6.37 (0.26) |
| Saturated Fat | 5.63 (0.07) | 7.83 (0.08)* | 6.36 (0.11)* | 6.11 (0.08)* | 6.87 (0.11)* | 6.02 (0.20) |
| Added Sugar | 6.96 (0.07) | 8.56 (0.10)* | 6.12 (0.08)* | 6.98 (0.12) | 6.97 (0.10) | 6.33 (0.24)* |
| Total Score | 54.32 (0.34) | 59.80 (0.53)* | 51.42 (0.45)* | 51.99 (0.44)* | 55.26 (0.47) | 52.00 (1.10)* |
SE = Standard Error.
One way analysis of variance was used to explore bivariate association with race/ethnicity, adjusting for the NHANES complex sampling design. Race/ethnicity was significant for all of the dietary variables. An asterisk (*) indicates significant mean difference from White.
Health Eating Index: The possible ranges of all components are 0–10 except total vegetable, green/bean, fruit, whole fruit, sea plant, and protein whose range is 0–5. The total score ranges 0–100. A higher score of the Health Eating Index component indicates a better quality of the dietary component.
Footnotes
Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of a an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.
Data collection for the National Health and Nutrition Examination Survey was approved by the National Center for Health Statistics Research Ethics Review Board. Analysis of de-identified data from the survey is exempt from federal regulations for the protection of human research participants (Protocol #2011–17). An individual investigator utilizing the publicly available NHANES data do not need to file the institution internal review board (IRB).
Conflicts of Interest/competing interests. None of the authors identify any conflict of interest and has no relevant or material financial interest related to this paper.
Availability of data and material. The data are available in the National Health and Nutrition Examination Survey website, https://www.cdc.gov/nchs/nhanes/index.htm.
Code availability. The corresponding author can provide the analysis R codes via an email request.
REFERENCES
- 1.Lépine JP, Briley M, The increasing burden of depression. Neuropsychiatr Dis Treat. 2011;7(Suppl 1):3–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.National Institute of Mental Health. Major depression. 2019. https://www.nimh.nih.gov/health/statistics/major-depression.shtml. Accessed 9 Oct 2019.
- 3.Kessler RC, Chiu WT, Demler O, Merikangas KR, Walters EE. Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry. 2005. June;62(6):617–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.World Health Organization. Depression. 2018. https://www.who.int/news-room/fact-sheets/detail/depression. Accessed 23 Oct 2019.
- 5.Martínez-González MA, Sánchez-Villegas A. Food patterns and the prevention of depression. Proc Nutr Soc. 2016. May;75(2):139–46. [DOI] [PubMed] [Google Scholar]
- 6.Hawton K, Casañas I Comabella C, Haw C, Saunders K. Risk factors for suicide in individuals with depression: a systematic review. J Affect Disord. 2013. May;147(1–3):17–28. [DOI] [PubMed] [Google Scholar]
- 7.Jacka FN, Cherbuin N, Anstey KJ, Butterworth P. Does reverse causality explain the relationship between diet and depression? J Affect Disord. 2015. April 1;175:248–50. [DOI] [PubMed] [Google Scholar]
- 8.De Mello MT, Lemos Vde A, Antunes HK Bitterncourt L, Santos-Silva R, Tufik S. Relationship between physical activity and depression and anxiety symptoms: a population study. J Affect Disord. 2013. July;149(1–3):241–6. [DOI] [PubMed] [Google Scholar]
- 9.Phillips WT, Kiernan M, King AC. Physical activity as a nonpharmacological treatment for depression: a review. Complement Health Pract Rev. 2003. April;8(2):139–52. [Google Scholar]
- 10.Mammen G, Faulkner G. Physical activity and the prevention of depression: a systematic review of prospective studies. Am J Prev Med. 2013. November;45(5): 649–57. [DOI] [PubMed] [Google Scholar]
- 11.Riolo SA, Nguyen TA, Greden JF, King CA. Prevalence of depression by race/ethnicity: findings from the National Health and Nutrition Examination Survey III. Am J Public Health. 2005. June;95(6):998–1000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Steffens DC, Fisher GG, Langa KM, Potter GG, Plassman BL. Prevalence of depression among older Americans: the Aging, Demographics and Memory Study. Int Psychogeriatr. 2009. October;21(5):879–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kalibatseva Z, Leong FT. Depression among Asian Americans: review and recommendations. Depress Res Treat, 2011;2011:320902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Breslau J, Kendler KS, Su M, Gaxiola-Aguilar S, Kessler RC. Lifetime risk and persistence of psychiatric disorders across ethnic groups in the United States. Psychol Med. 2005. March;35(3):317–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Huang B, Grant BF, Dawson DA, Stinson FS, Chou SP, Saha TD, Doldstein RB, Smith SM, Ruan WJ, Pickering RP. Race-ethnicity and the prevalence and co-occurrence of Diagnostic and Statistical Manual of Mental Disorders, alcohol and drug use disorders and Axis I and II disorders: United States, 2001 to 2002. Compr Psychiatry. 2006. Jul-Aug;47(4):252–7. [DOI] [PubMed] [Google Scholar]
- 16.Kisch J, Leino EV, Silverman MM. Aspects of suicidal behavior, depression, and treatment in college students: results from the spring 2000 national college health assessment survey. Suicide Life Threat Behav. 2005. February;35(1):3–13. [DOI] [PubMed] [Google Scholar]
- 17.Brown A U.S. Hispanic and Asian populations growing, but for different reasons. 2014. https://www.pewresearch.org/fact-tank/2014/06/26/u-s-hispanic-and-asian-populations-growing-but-for-different-reasons/. Accessed 23 Oct 2019.
- 18.Beydoun MA, Wang Y. Pathways linking socioeconomic status to obesity through depression and lifestyle factors among young US adults. J Affect Disord. 2010. June;123(1–3):52–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Givens JL, Houston TK, Van Voorhees BW, Ford DE, Cooper LA. Ethnicity and preferences for depression treatment. Gen Hosp Psychiatry. 2007. May-Jun;29(3):182–91. [DOI] [PubMed] [Google Scholar]
- 20.Wang PS, Lane M, Olfson M, Pincus HA, Wells KB, Kessler RC. Twelve-month use of mental health services in the United States: results from the National Comorbidity Survey Replication. Arch Gen Psychiatry. 2005. June;62(6):629–40. [DOI] [PubMed] [Google Scholar]
- 21.Bauldry S, Szaflarski M. Immigrant-based disparities in mental health care utilization. Socius. 2017. January;3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Dash S, Clarke G, Berk M, Jacka FN. The gut microbiome and diet in psychiatry: focus on depression. Curr Opin Psychiatry. 2015. January;28(1):1–6. [DOI] [PubMed] [Google Scholar]
- 23.Lim E Gandhi K, Siriwardhana C, Davis J, Chen JJ. Racial and ethnic differences in mental health service utilization among the Hawaii medicaid population. J Ment Health. 2019. October;28(5):536–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001. September;16(9):606–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kennedy ET, Ohis J, Carlson S, Fleming K. The healthy eating index: design and applications. J Am Diet Assoc. 1995. October;95(10):1103–8. [DOI] [PubMed] [Google Scholar]
- 26.National Cancer Institute. Healthy Eating Index. 2019. https://epi.grants.cancer.gov/hei/sas-code.html. Accessed 31 Jan 2020.
- 27.Reedy J, Lerman JL, Krebs-Smith SM, Kirkpatrick SI, Pannucci TE, Wilson MM, Subar AF, Kahle LL, Tooze JA. Evaluation of the Healthy Eating Index-2015. J Acad Nutr Diet. 2018. September;118(9):1622–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Armstrong T, Bull F. Development of the world health organization global physical activity questionnaire (GPAQ). J Public Health. 2006. April 1;14(2):66–70. [Google Scholar]
- 29.Centers for Disease Control and Prevention. National Health and Nutrition Examination Survey. Physical activity (PAQ_I). 2017. https://wwwn.cdc.gov/Nchs/Nhanes/2015-2016/PAQ_I.htm. Accessed 9 Oct 2019.
- 30.Kim WK, Shin D, Song WO. Are dietary patterns associated with depression in U.S. adults? J Med Food. 2016. November;19(11):1074–84. [DOI] [PubMed] [Google Scholar]
- 31.Jacka FN, O’Neil A, Opie R, Itsiopoulos C, Cotton S, Mohebbi M, Castle D, Dash S, Mihalopoulos C, Chatterton ML, Brazionis L, Dean OM, Hodge AM, Berk M. A randomised controlled trial of dietary improvement for adults with major depression (the ‘SMILES’trial). BMC Med. 2017. January 30;15(1):23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Cohut M Mediterranean diet may protect against depression symptoms. 2019. https://www.medicalnewstoday.com/articles/325240.php. Accessed 30 Oct 2019.
- 33.Sánchez-Villegas A, Delgado-Rodriguez M, Alonso A, Schlatter J, Lahortiga F, Serra Majem L, Martinez-González MA. Association of the Mediterranean dietary pattern with the incidence of depression: the Seguimiento Universidad de Navarra/University of Navarra follow-up (SUN) cohort. Arch Gen Psychiatry. 2009. October;66(10):1090–8. [DOI] [PubMed] [Google Scholar]
- 34.Dai J, Miller AH, Bremner JD, Goldberg J, Jones L, Shalenberger L, Buckha R, Murrah NV, Veldear E, Wilson PW, Vaccarino V. Adherence to the Mediterranean diet is inversely associated with circulating interleukin-6 among middle-aged men: a twin study. Circulation. 2008. January 15;117(2):169–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Lamers F, Milaneschi Y, Smit JH, Schoevers RA, Wittenberg G, Penninx BWJH. Longitudinal association between depression and inflammatory markers: results from the Netherlands Study of Depression and Anxiety. Biol Psychiatry. 2019. May 15;85(10):829–37. [DOI] [PubMed] [Google Scholar]
- 36.Whitaker KM, Sharpe PA, Wilcox S, Hutto BE. Depressive symptoms are associated with dietary intake but not physical activity among overweight and obese women from disadvantaged neighborhoods. Nutr Res. 2014. April;34(4):294–301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Payne ME, Steck SE, George rR, Steffens DC. Fruit, vegetable, and antioxidant intakes are lower in older adults with depression. J Acad Nutr Diet. 2012. December;112(12):2022–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Arab L, Guo R, Elashoff D. Lower depression scores among walnut consumers in NHANES. Nutrients. 2019. January 26;11(2):e275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Madison A, Kiecolt-Glaser JK. Stress, depression, diet, and the gut microbiota: Human–bacteria interactions at the core of psychoneuroimmunology and nutrition. Curr Opin Behav Sci. 2019;28:105–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Brown HE, Pearson N, Braithwaite RE, Brown WJ, Biddle SJ. Physical activity interventions and depression in children and adolescents. Sports Med. 2013. March;43(3):195–206. [DOI] [PubMed] [Google Scholar]
- 41.Herring MP, Puetz TW, O’Connor PJ, Dishman RK. Effect of exercise training on depressive symptoms among patients with a chronic illness: a systematic review and meta-analysis of randomized controlled trials. Arch Intern Med. 2012. January 23;172(2):101–11. [DOI] [PubMed] [Google Scholar]
- 42.Galenkamp H, Stronks K, Snijder MB, Derks EM. Measurement invariance testing of the PHQ-9 in a multi-ethnic population in Europe: the HELIUS study. BMC Psychiatry. 2017. October 24;17(1):349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Baas KD, Cramer AO, Koeter MW, van de Lisdonk EH, van Weert HC, Schene AH. Measurement invariance with respect to ethnicity of the Patient Health Questionnaire-9 (PHQ-9). J Affect Disord. 2011. March;129(1–3):229–35. [DOI] [PubMed] [Google Scholar]
- 44.Huang FY, Chung H, Kroenke K, Delucchi KL, Spitzer RL. Using the patient health questionnaire-9 to measure depression among racially and ethnically diverse primary care patients. J Gen Intern Med. 2006. June;21(6):547–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Sohail Z, Bailey RK, Richie WD. Misconceptions of depression in african americans. Front Psychiatry. 2014. June 20;5:65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Mulsant BH, Ganguli M, Seaberg EC. The relationship between self-rated health and depressive symptoms in an epidemiological sample of community-dwelling older adults. J Am Geriatr Soc. 1997. August;45(8):954–8. [DOI] [PubMed] [Google Scholar]
- 47.Nutt D, Wilson S, Paterson L. Sleep disorders as core symptoms of depression. Dialogues Clin Neurosci. 2008;10(3):329–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Winfield BH, Mizuno T, Beaudoin CE. Confucianism, collectivism and constitutions: Press systems in China and Japan. Comm L & Pol’y. 2000;5(3):323–47. [Google Scholar]
- 49.Fortuna LR, Porche MV, Alegria M. Political violence, psychosocial trauma, and the context of mental health services use among immigrant Latinos in the United States. Ethn Health. 2008;13(5):435–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Healthcare.Gov. Health benefits & coverage: Mental health & substance abuse coverage. 2019. https://www.healthcare.gov/coverage/mental-health-substance-abuse-coverage/. Accessed 31 Oct 2019.
- 51.Hann CS, Rock CL, King I, Drewnowski A. Validation of the Healthy Eating Index with use of plasma biomarkers in a clinical sample of women. Am J Clin Nutr. 2001;74(4):479–86. [DOI] [PubMed] [Google Scholar]
- 52.Adjoian TK, Firestone MJ, Eisenhower D, Yi SS. Validation of self-rated overall diet quality by Healthy Eating Index-2010 score among New York City adults, 2013. Prev Med Rep. 2016;3:127–31. [DOI] [PMC free article] [PubMed] [Google Scholar]