- Journal List
- HHS Author Manuscripts
- PMC7891298
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsem*nt of, or agreement with, the contents by NLM or the National Institutes of Health.
Learn more: PMC Disclaimer | PMC Copyright Notice
Ann Epidemiol. Author manuscript; available in PMC 2021 Feb 18.
Published in final edited form as:
Ann Epidemiol. 2020 Mar; 43: 25–31.e1.
Published online 2020 Feb 8. doi:10.1016/j.annepidem.2020.02.001
PMCID: PMC7891298
NIHMSID: NIHMS1636826
PMID: 32147320
Catherine K. Ettman,a,b,* Gregory H. Cohen, MPhil, MSW,c and Sandro Galea, MD, DrPHa
Author information Copyright and License information PMC Disclaimer
The publisher's final edited version of this article is available at Ann Epidemiol
Associated Data
- Supplementary Materials
Abstract
Purpose:
The aim of the study was to assess the relation between family wealth and depression in U.S. adults.
Methods:
Participants were 5134 members of the 2015–2016 National Health and Nutrition Examination Survey who were aged 18 years or older and completed the depression screener. Using the Patient Health Questionnaire and household demographics interview data, we calculate the adjusted odds of depressive symptoms for persons with low relative to high family savings, using multivariable logistic regression. We estimate predicted probabilities of having depressive symptoms for low and high family savings groups at low, middle, and high family income categories.
Results:
Overall, 57.4% of the total weighted population had low family savings (below $20,000), and 23.7% of the weighted population had depressive symptoms. Persons with low family savings had 1.49 times higher odds (95% confidence interval, 1.01–2.21) of having depressive symptoms than persons with high family savings, controlling for gender, age, race, education, marital status, family size, and family income. Predicted probabilities of depressive symptoms were higher for low family savings groups than high family savings groups at every income level.
Conclusions:
Family wealth is associated with lower prevalence of current depressive symptoms in U.S. adults. Wealth may be an important determinant of population mental health, separate and independent from income.
Keywords: Mental health, Depression, Socioeconomic factors, Healthcare disparities, Stress, Psychological, Economics
Introduction
Depression is a leading cause of disability in the U.S. In total, 17.3 million adults experienced a major depressive disorder in the United States in 2017 [1]. From 2013 to 2016, 8.1% of adults aged older than 20 years in the United States experienced depression in a given 2-week period [2]. Depression affects not only the persons with symptoms but also their family, friends, and coworkers. Depression costs the U.S. $210 billion annually [3]. Additional costs include increased absenteeism in the workplace, decreased marital and parental functioning, and reduced earnings [4].
Mental health is, in part, shaped by the social and economic conditions in which people live [5]. For example, depression is associated with low educational attainment, material disadvantage, and marital status [6,7]. It is also well documented that low income is a risk factor for depression [2,8–11]. Higher income may protect against depression by reducing stressors and providing social resources; financial resources may also be used to pay for treatment for existing cases of depression.
Although income captures current cash flow, wealth—representing lifetime accumulation of resources—may play a different role than income. Wealth has been associated with improved life expectancy and reduced disability [12,13]. Wealth may both protect persons from the effects of low income and, in and of itself, represent a separate pathway that protects persons from adversities and attendant poor health. For example, wealth may allow a family to purchase a home or to move to a safer neighborhood. Wealth may also insulate people from the psychological stressors of financial strain more than income [14]. A study of twins showed that perceived wealth, or subjective assessment of one’s financial situation, may mediate the relationship between objective wealth and life satisfaction [15]. Having assets beyond income can provide financial stability in the face of unexpected shocks; in addition, wealth may increase social status separate from income, thereby providing more pathways to better health.
Wealth may, therefore, be an important determinant of depression, separate from the role that income plays in shaping the risk of depression. There is, however, very little known about the connection between family wealth and mental health. First, of the studies on wealth, much evidence uses wealth as an ecologic variable instead of looking at a person’s access to individual or family wealth. For example, studies have looked at country-level wealth using gross domestic product, Gini index, or a wealth index using variants of the market values of a standard set of household goods [13,16,17]. At the country level, increased perceived wealth is associated with reduced mental illness [16]. Second, there are many types of wealth, making it challenging to record. There are few formal mechanisms to document individual or family wealth (unlike income, which is documented through annual taxes), persons may be unwilling to share the extent of their full assets, and it is difficult to add up the combination of liquid and nonliquid assets—such as property—contributing to a family’s wealth. In the one study of which we are aware that looks at wealth at the family level and individual mental health, Lê-Scherban et al. found that adolescents who were raised in families with greater household wealth experienced lower prevalence of serious psychological distress compared with adolescents from families with wealth below median levels [18]. Lê-Scherban’s study included 2060 young adults aged 18–27 years in 2005–2011 in the United States. There have been few studies that look at individual-level current savings and current depression in adults.
Despite growing sophistication in our collective understanding of economic determinants of health, the relationship between wealth and mental health remains a missing piece. We took advantage of a nationally representative survey to assess the relationship between family wealth and depressive symptoms in the United States.
Methods
Sample
The National Health and Nutrition Examination Survey (NHANES) is a series of cross-sectional surveys that are representative of the U.S. population, conducted by the U.S. Government annually. Data are collected from household interviews and medical examinations conducted in the Medical Examination Centers (MECs) across 30 geographic sites. In 2015–2016, 15,327 persons were identified by NHANES from 30 survey locations across the country to participate in the survey. Of those, 9971 completed the household interview (where the demographic and income information are collected), and 9544 were examined in the MEC. Of those examined in the MEC, participants aged younger than 12 years and participants requiring a proxy were not eligible for the depression screener because of the sensitive nature of questions asked. In addition, some persons eligible for the depression screener who participated in the MEC did not take the depression screener because of time constraints and refusals [19]. We excluded from our sample any persons aged younger than 18 years (n = 3979) and persons with any missing depression data (n = 858). The final sample population included 5134 participants.
Study constructs
Depression
Depression was defined using the Patient Health Questionnaire (PHQ-9), a depression screener that incorporates the Diagnostic and Statistical Manual of Mental Disorders-IV depression diagnostic criteria [20]. Questionnaires were distributed via a computer-assisted personal interview. Questions included, “Over the last 2 weeks, how often have you been bothered by the following problems: little interest or pleasure in doing things?” Response categories included “Not at all,” “Several days,” “More than half the days,” and “Nearly every day.” Answers were coded values of 0, 1, 2, or 3, respectively. Answers of “Refused” or “Don’t know” were coded as missing values. All answers were added into a composite score for each person, ranging from 0 to 27. For the purpose of this analysis, we used a cutoff score of 5 or more, consistent with others, as a threshold for depressive symptoms [21].
Wealth
Wealth was defined as a binary variable (high/low) with a code of “high” referring to having more than $20,000 in family savings. Participants were asked, “Do {you/NAMES OF OTHER FAMILY/you and NAMES OF FAMILY MEMBERS} have more than $20,000 in savings at this time? Please include money in your checking accounts.” Family savings included cash, checking account, saving accounts, certificates of deposit, retirement accounts (such as IRAs, 401K, etc.), stocks, and bonds mutual funds.
Covariables of interest
We controlled for the following covariables: race/ethnicity, recoded into five categories (1 = non-Hispanic white, 2 = non-Hispanic black, 3 = Hispanic, 4 = non-Hispanic Asian, and 5 = other, including multiracial); gender (coded as “female” or “male”); marital status (“married,” “widowed, divorced, or separated,” “never married,” or “living with partner”); family income (“Under $20,000,” “$20,000–$75,000,” and “$75,000+”); and age. Age was coded as a categorical variable, consistent with NHANES analytical guidelines to optimize survey weights, with categories of 18–39, 40–59, and 60+. Family size was defined as “a group of people related by birth, marriage, or adoption and residing together.” Families with more than seven people were labeled as “7 or more” to preserve the confidentiality of participants.
Analysis
Weights accounting for the complex survey design of NHANES were applied in all analyses. We used the Mobile Exam Center weights, consistent with NHANES statistical guidelines. First, we used two-tailed Pearson χ2 analysis to examine the bivariable relationship between the covariables of interest and depression and the covariables of interest and savings, using the svy: tabulate command in STATA. Percentages reported are weighted responses. Second, we constructed multiple logistic regression models to test the odds of depressive symptoms for those with high and low savings, controlling for demographic characteristics, using the svy: logistic command in STATA. We calculated odds ratios (ORs) and 95% confidence intervals (CIs) for the whole sample and stratified by income level. Third, we calculated predicted probabilities of depressive symptoms by wealth across levels of income, adjusted for the average values of demographic covariables, and by income across levels of wealth. We used STATA statistical software, version 15.1, (StataCorp LP, College Station, TX) to analyze all data [22].
Sensitivity analysis
We conducted two sensitivity analyses to confirm the robustness of results. First, we reran all multivariable logistic regression models across various cutoffs of depression, at 5, 10, and 15 to assess whether the relationship holds across depression severity, consistent with validated PHQ-9 cutoffs [20]. Second, we created four scenarios to test out varying data structures that underlie missingness for our main exposure variable to see whether the relationship between depression and savings holds when values are imputed for the missing savings values using two exposure outcome patterns: pattern 1 sets those with depressive symptoms to high savings and those without depressive symptoms to low savings, whereas pattern 2 sets those with depressive symptoms to low savings and those without depressive symptoms to high savings. In scenario 1, we randomly assigned pattern 1%–50% of participants with missing values, and pattern 2 to the other 50%. In scenario 2, we assigned pattern 1%–75% of participants and pattern 2 to the other 25%. In scenario 3, we assigned pattern 1%–85% of participants and pattern 2 to the other 15%. In scenario 4, we assigned pattern 1%–100% of participants with missing values.
Results
Of the 5134 persons in the sample, 1309 participants had current depressive symptoms and 3511 had low family savings (less than $20,000). Overall, 23.7% of the weighted population had depressive symptoms. Table 1 shows the demographic characteristics of the survey sample weighted to the U.S. population; it also shows the bivariable relationships between these demographics and savings and depressive symptoms. Variables associated with depressive symptoms included gender, education, marital status, family income, and family savings (P < .001). The prevalence of depressive symptoms was higher in non-Hispanic blacks (25.2%), Hispanics (24.7%), and other (33.5%) than in non-Hispanic whites (23.4%) or non-Hispanic Asians (15.8%). The prevalence of current depressive symptoms was higher in women (28.1%) than men (19.2%). More education was associated with less depression. The prevalence of depressive symptoms was greater among participants with high school or less (29.2%) than among participants with college or more (20.6%). The prevalence of depressive symptoms was lowest among participants who were married (18.8%) and highest among participants who were widowed, divorced, or separated (34.2%). Having lower income was associated with higher prevalence of depressive symptoms. The prevalence of depressive symptoms was 39.3% for participants with family income under $20,000, 25.5% for participants with family income from $20,000–$75,000, and 14.9% for participants with family income greater than $75,000. The prevalence of current depressive symptoms was greater among participants with low family savings (29.1%) than among those with high family savings (15.6%). Age was not significantly associated with the prevalence of depressive symptoms.
Table 1
Demographic characteristics in 2015–2016 NHANES study sample (n = 5134) and bivariable analyses with current depressive symptoms and with savings
| Characteristics | N | Weighted % from sample | Low savings | Depressive symptoms | ||||
|---|---|---|---|---|---|---|---|---|
| n | %* | P† | n | %* | P† | |||
| Gender | ||||||||
| Male | 2508 | 48.7% | 1688 | 56.9 | .671 | 537 | 19.2% | <.001 |
| Female | 2626 | 51.3% | 1823 | 57.9 | – | 772 | 28.1% | – |
| Age | ||||||||
| 18–39 | 1867 | 36.9% | 1418 | 70.4 | <.001 | 457 | 23.8% | .425 |
| 40–59 | 1602 | 35.2% | 1075 | 56.1 | – | 419 | 25.2% | – |
| 60+ | 1665 | 27.9% | 1018 | 41.9 | – | 433 | 21.8% | – |
| Race | ||||||||
| Non-Hispanic white | 1710 | 64.8% | 952 | 48.3 | <.001 | 460 | 23.4% | .025 |
| Non-Hispanic black | 1096 | 11.2% | 850 | 77.0 | – | 278 | 25.2% | – |
| Hispanic | 1590 | 15.2% | 1292 | 82.0 | – | 419 | 24.7% | – |
| Non-Hispanic Asian | 541 | 5.1% | 273 | 47.7 | – | 88 | 15.8% | – |
| Other | 197 | 3.7% | 144 | 69.7 | – | 64 | 33.5% | – |
| Education | ||||||||
| High school or less | 2452 | 36.4% | 1966 | 73.6 | <.001 | 722 | 29.2% | <.001 |
| College or more | 2681 | 63.6% | 1544 | 48.1 | – | 586 | 20.6% | – |
| Missing | 1 | – | 1 | – | 1 | – | ||
| Marital status | ||||||||
| Married | 2460 | 53.1% | 1491 | 47.1 | <.001 | 510 | 18.8% | <.001 |
| Widowed, divorced, or separated | 1054 | 17.6% | 777 | 66.2 | – | 350 | 34.2 | – |
| Never married | 898 | 17.2% | 668 | 69.8 | – | 266 | 27.2 | – |
| Living with partner | 475 | 9.6% | 390 | 74.6 | – | 113 | 25.3 | – |
| Missing | 247 | 2.5% | 185 | 64.8 | – | 70 | 25.8 | – |
| Family income | ||||||||
| Under $20,000 | 1181 | 15.7% | 1091 | 88.3 | <.001 | 444 | 39.3 | <.001 |
| $20,000-$75,000 | 2332 | 43.5% | 1865 | 73.4 | – | 580 | 25.5 | – |
| $75,000+ | 1191 | 34.1% | 442 | 29.8 | – | 194 | 14.9 | – |
| Missing | 430 | 6.7% | 113 | 21.9 | – | 91 | 21.3 | – |
| Savings | ||||||||
| Low | 3511 | 57.4% | – | – | – | 210 | 29.1 | <.001 |
| High | 1257 | 37.2% | – | – | – | 1025 | 15.6 | – |
| Missing | 366 | 5.4% | – | – | – | 74 | 23.3 | – |
| Depressive symptoms | ||||||||
| Yes | 1309 | 23.7% | 1025 | 70.3 | <.001 | – | – | – |
| No | 3825 | 76.3% | 2486 | 53.4 | – | – | – | |
| – | Mean* | – | Mean* | – | – | Mean* | – | |
| Family size | – | 2.9 | – | 3.0 | – | – | 2.8 | – |
Open in a separate window
*Weighted percentages.
†Two-tailed χ2 analysis done for significance testing at 95% level. χ2 tests show differences of distributions across characteristics.
Overall, 57.4% of the total weighted population had low family savings (below $20,000). Variables associated with having low family savings included age, race, education, marital status, family income, and current depressive symptoms (P < .001). The prevalence of having low family savings was greater among participants aged 18–39 years (70.4%) than participants aged 40–59 (56.1%) and among participants aged 60+ years (41.9%). The prevalence of having low family savings by race varied as follows: Non-Hispanic white (48.3%), non-Hispanic black (77.0%), Hispanic (82.0%), non-Hispanic Asian (47.7%), and other (69.7%). Having less education was associated with a higher prevalence of having low family income: among participants with high school education or less, 73.6% had low family savings; among participants with college or more education, 48.1% had low family savings. Married persons had a lower prevalence of having low savings (47.1%) than persons who were widowed, divorced, or separated (66.2%), never married (69.8%), or living with a partner (74.6%). Having lower family income was associated with greater prevalence of low family savings. Of persons with a family income under $20,000, 88.3% had low family savings; of persons with a family income of $20,000–$75,000, 73.4% had low family savings; of persons with a family income greater than $75,000, 29.8% had low family savings. Of persons with depressive symptoms, 70.3% had low family savings. Of persons without depressive symptoms, 53.4% had low family savings.
Table 2 shows the multivariable models describing the relations between savings and depressive symptoms. Model I shows the simple logistic regression with gender and depression. In Models II to V, we add the additional covariables of age, race, education, and marital status in a step-wise fashion. Models VI and VII show these variables plus either family income or family savings, controlling for family size. Model VI shows the association of income and depressive symptoms without controlling for savings: OR 3.32 (95% CI, 2.40–4.61) for family income under $20,000, and OR 1.86 (95% CI, 1.40–2.46) for family income from $20,000–$75,000 relative to family income $75,000+. Model VII shows the association of family savings and depressive symptoms without controlling for income: OR 1.92 (95% CI, 1.37–2.68) for low relative to high family savings. Model VIII shows our final model with gender, age, race, education, marital status, family size, family income, and family savings. Variables associated with the prevalence of depressive symptoms in the final multivariable model were gender, education, marital status, family income, and family savings (Table 2). In the final multivariable model, persons with low family savings had 1.49 times higher odds of having depressive symptoms than persons with high family savings (above $20,000; 95% CI, 1.01–2.21), controlling for gender, age, race, education, marital status, family size, and family income. In the final multivariable model, persons with family income less than $20,000 had 2.74 times the odds of having depression than persons with family income greater than $75,000 (95% CI, 1.87–4.01), controlling for gender, age, race, education, marital status, family size, and family savings. In the models stratified across income (not shown), the association between depression and family savings was strongest for the middle-income group making between $20,000 and $75,000 in annual family income (OR 2.03; 95% CI, 1.12–3.70). The association was not significant for the low-income group (OR 1.45; 95% CI, 0.60–3.51) or for the high-income group (OR 1.02; 95% CI, 0.66–1.59).
Table 2
Multivariable models describing the relation between savings and depressive symptoms in 2015–2016 NHANES study sample (n = 5134)
| Characteristics | Model I | Model II | Model III | Model IV | Model V | Model VI | Model VII | Model VIII | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | |
| Gender | ||||||||||||||||
| Male | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| Female | 1.64 | 1.35–2.01 | 1.7 | 1.35–2.02 | 1.66 | 1.36–2.02 | 1.72 | 1.42–2.10 | 1.65 | 1.36–2.00 | 1.64 | 1.33–2.02 | 1.66 | 1.35–2.05 | 1.64 | 1.34–2.02 |
| Age | ||||||||||||||||
| 18–39 | – | – | 1.2 | 0.84–1.57 | 1.14 | 0.83–1.57 | 1.18 | 0.87–1.61 | 1.21 | 0.91–1.62 | 1.29 | 0.95–1.74 | 1.42 | 1.10–1.85 | 1.19 | 0.87–1.62 |
| 40–59 | – | – | 1.2 | 0.92–1.65 | 1.23 | 0.91–1.66 | 1.31 | 0.99–1.73 | 1.42 | 1.08–1.87 | 1.65 | 1.25–2.17 | 1.12 | 0.82–1.52 | 1.55 | 1.18–2.04 |
| 60+ | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| Race | ||||||||||||||||
| Non-Hispanic White | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| Non-Hispanic Black | – | – | – | – | 1.07 | 0.78–1.48 | 0.99 | 0.72–1.36 | 0.88 | 0.64–1.20 | 0.76 | 0.56–1.03 | 0.79 | 0.57–1.10 | 0.73 | 0.53–1.00 |
| Hispanic | – | – | – | – | 1.06 | 0.85–1.33 | 0.88 | 0.68–1.14 | 0.86 | 0.67–1.10 | 0.76 | 0.58–0.99 | 0.81 | 0.60–1.09 | 0.71 | 0.53–0.96 |
| Non-Hispanic Asian | – | – | – | – | 0.60 | 0.39–0.92 | 0.6 | 0.40–0.90 | 0.6 | 0.40–0.90 | 0.61 | 0.37–1.01 | 0.66 | 0.40–1.10 | 0.62 | 0.35–1.08 |
| Other | – | – | – | – | 1.65 | 1.05–2.61 | 1.64 | 1.06–2.55 | 1.48 | 0.95–2.30 | 1.37 | 0.82–2.28 | 1.39 | 0.87–2.23 | 1.33 | 0.80–2.20 |
| Education | ||||||||||||||||
| High school or less | – | – | – | – | – | – | 1.71 | 1.40–2.10 | 1.65 | 1.33–2.04 | 1.36 | 1.10–1.68 | 1.49 | 1.23–1.81 | 1.32 | 1.08–1.62 |
| College or more | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| Marital status | ||||||||||||||||
| Married | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| Widowed, divorced, or separated | – | – | – | – | – | – | 2.02 | 1.70–2.42 | 1.57 | 1.32–1.87 | 1.78 | 1.47–2.16 | 1.51 | 1.25–1.82 | ||
| Never married | – | – | – | – | – | – | – | – | 1.62 | 1.35–1.94 | 1.33 | 1.10–1.62 | 1.5 | 1.19–1.87 | 1.32 | 1.07–1.63 |
| Living with partner | – | – | – | – | – | – | – | – | 1.39 | 1.07–1.81 | 1.17 | 0.95–1.45 | 1.28 | 0.99–1.65 | 1.15 | 0.94–1.40 |
| Family income | ||||||||||||||||
| Under $20,000 | – | – | – | – | – | – | – | – | – | – | 3.32 | 2.40–4.61 | – | – | 2.74 | 1.87–4.01 |
| $20,000-$75,000 | – | – | – | – | – | – | – | – | – | – | 1.86 | 1.40–2.46 | – | – | 1.61 | 1.14–2.27 |
| $75,000 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| Family savings | ||||||||||||||||
| Low | – | – | – | – | – | – | – | – | – | – | – | – | 1.92 | 1.37–2.68 | 1.49 | 1.01–2.21 |
| High | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| Family size | – | – | – | – | – | – | – | – | – | – | 1.04 | 0.99–1.10 | 0.98 | 0.93–1.03 | 1.03 | 0.97–1.09 |
Open in a separate window
Weighted estimates used in all models.
Figure 1 shows covariable-adjusted predicted probabilities for depressive symptoms for persons with low family savings and with high family savings at low, middle, and high family income levels. The predicted probability of depressive symptoms for persons with the lowest level of family income (below $20,000) was 0.38 (95% CI, 0.34–0.43) for persons with low family savings and 0.30 (95% CI, 0.22–0.38) for persons with high family savings. The predicted probability of depression for persons with the middle level of family income ($20,000–$75,000) was 0.27 (95% CI, 0.24–0.30) for persons with low family savings and 0.20 (95% CI, 0.14–0.26) for persons with high family savings. The predicted probability of depression for persons with the highest level of family income ($75,000+) was 0.19 (95% CI, 0.14–0.24) for persons with low family savings and 0.14 (95% CI, 0.10–0.17) for persons with high family savings. In the models stratified across income (not shown), higher savings groups had lower predicted probability of depression at every income level; the difference in predicted probability between income groups for low and high savings was greater for the low family income (10.4% difference) and middle family income (11.4% difference) than for the high family income group (2.5% difference.)
Fig. 1.
Predicted probability of depressive symptoms by savings at different levels of family income. Final regression model used, controlling for gender, age, race, education, family size, and marital status.
Sensitivity analyses
To assess the robustness of the association between wealth and depression, we conducted several sensitivity analyses under different cutoffs of depression and under different conditions of missingness.
In Table S1, we conduct the same analyses at cutoffs of 10 and 15, finding effect sizes for moderate (PHQ-9 cutoff of 10; OR, 1.59 [95% CI, 0.80–3.19]) and moderately severe depression (PHQ-9 cutoff of 15; OR, 3.31 [95% CI, 1.07–10.22]).
Three hundred sixty-six participants were missing data on family savings. To assess the impact of missingness of savings on the relationship between savings and depression, we conducted several deterministic imputation sensitivity analyses under varying assumptions of underlying data. In each of the scenarios reported in Table S2 and described further in the e-appendix, the point estimate of the effect of savings on depression is similar to the effect observed using only the complete data. Point estimates for imputation scenarios ranged from 1.38 to 1.46 across the varying assumptions of missingness.
Discussion
Using data representative of the U.S. population, we found that adults with low family savings were more likely to have depressive symptoms than adults with high family savings. Controlling for gender, age, race, education, marital status, and family income, having low family savings was associated with 50% greater odds of depressive symptoms as persons with high family savings. The association with wealth and depressive symptoms remained when controlling for all levels of income; persons with low family savings showed higher predicted probabilities of depressive symptoms at low-, middle-, and high-income categories than persons with high family savings in the same income categories.
Our findings suggest that wealth may provide a buffer against depressive symptoms before and after adjusting for family income. In addition, our findings suggest that wealth may serve as a buffer against the ill effects of lower income on depressive symptoms. The underlying finding of the study, namely that lower wealth is associated with greater burden of depressive symptoms, is consistent with findings of other studies that have explored the issue of economic resources and mental health [18]. Our findings are consistent with Pool et al.’s findings that negative wealth shock is associated with elevated depressive symptoms among late middle-aged adults (OR, 1.50; 95% CI, 1.10–2.05) [23]. Other studies on financial strain have shown an association between increased perceived financial strain and reduced mental health [24–26]. Dijkstra-Kersten et al. found that mild or severe financial strain was associated with increased odds of depression, above and beyond income (OR, 1.68; 95% CI, 1.35–2.09; and OR, 3.88; 95% CI, 2.58–5.81) [14].
Having access to wealth comes with a set of resources that may help to protect persons from depression and the factors associated with it, such as financial strain; our findings suggest that there are pathways through which wealth affects depression that are independent of the pathways linking income and depression. Wealth can embed access to a set of resources that buffer against adversities more robustly than does income. For example, although high earning persons without wealth may still deal with substantial economic insecurity, persons with wealth may have less financial insecurity, buffering against the relation between other life stressors and depression. The relation between life stressors and depression has been well demonstrated [26–29].
It should be noted that 57.4% of Americans did not have more than $20,000 in family savings; thus, the majority of Americans may face a greater risk of depressive symptoms because of lack of assets. This builds on a growing public awareness of many Americans’ inability to cover emergency expenses [30]. Wealth can be used to cover any range of expenses, including medical expenses, home expenses, or other necessities that could lead to stressors and mental illness.
There were three limitations of note to this work. First, the study measures current depressive symptoms as measured by mental health status in the past 2 weeks before taking the survey. Given that these associations look at one point in time, it is possible that there is a bidirectional relationship between depressive symptoms and wealth, with depression affecting family wealth accumulation. Given the life course pathway of wealth accumulation, we think it is reasonable to observe, as we do here, that wealth is associated with depressive symptoms and is a potential risk factor for individual mental illness. Second, this study does not account for the way that wealth was accumulated. The effects of inherited versus earned wealth on depressive symptoms may also differ. Third, because of low sample size within stratified income groups, it is difficult to comment meaningfully on the relationship between savings and depression within the lowest income group.
Notwithstanding these limitations, our findings suggest, perhaps not surprisingly, that wealth may serve as a protective factor against depressive symptoms. Wealth may be a proxy for access to other resources that are positively associated with mental health, including access to adequate nutrition, living, working, and playing in safe and health-promoting environments and having financial security in the case of expected or unexpected expenses. The definition of wealth we chose is above or below $20,000. There is a large difference between a family that has $25,000 in savings and $2,500,000 in savings; future work may fruitfully explore how different levels of wealth influence mental health, shedding light on potential pathways that explain these associations. It is also possible that wealth has differential effects across income levels; future work may look to the connection between different levels of wealth and different levels of income. Importantly, while people earn income, wealth is often inherited or gifted over time, with greater possibility for perpetuation of resource gaps and subsequent health gaps. Therefore, from a population health point of view, low wealth represents the risk of an intergenerational transfer of poor health that cannot be mitigated by an individual’s own behaviors. This has implications both for how we understand the patterning of population health and the potential solutions that may improve health. Simply put, individual persons can do little to overcome generationally transmitted risk factors such as wealth, and it falls to policy solutions to mitigate their adverse health consequences. To the best of our knowledge, this is the first study that shows an association between family wealth and depressive symptoms for adults in the United States. Future research may explore the causal mechanisms that explain the effect of wealth on depression.
Supplementary Material
1
Click here to view.(23K, pdf)
Acknowledgments
C.K.E. worked on this project while funded by the National Institutes of Health T32 AG 23482–15.
References
[1] Major Depression. NIMH mental health information statistics. https://www.nimh.nih.gov/health/statistics/major-depression.shtml. [Accessed 19 July 2019].
[2] Brody D, Pratt L, Hughes J. Prevalence of depression among adults aged 20 and over: United States, 2013–2016. U.S. Department of Health and Human Services Centers for Disease Control and Prevention National Center for Health statistics; Report No.: 303. https://www.cdc.gov/nchs/products/databriefs/db303.htm. [Accessed 19 July 2019]. [Google Scholar]
[3] Greenberg PE, Fournier AA, Sisitsky T, Pike CT, Kessler RC. The economic burden of adults with major depressive disorder in the United States (2005 and 2010). J Clin Psychiatry2015;76(02):155–62. [PubMed] [Google Scholar]
[4] Kessler RC. The costs of depression. Psychiatr Clin North Am2012;35(1):1–14. [PMC free article] [PubMed] [Google Scholar]
[5] Allen J, Marmot M, World Health Organization, Fundação Calouste Gulbenkian. Social determinants of mental health [Online], http://apps.who.int/iris/bitstream/10665/112828/1/9789241506809_eng.pdf?ua=1. [Accessed 20 July 2019].
[6] St John PD, Montgomery PR. Marital status, partner satisfaction, and depressive symptoms in older men and women. Can J Psychiatry2009;54(7): 487–92. [PubMed] [Google Scholar]
[7] Ruiz M, Hu Y, Martikainen P, Bobak M. Life course socioeconomic position and incidence of mid–late life depression in China and England: a comparative analysis of CHARLS and ELSA. J Epidemiol Community Health2019;73(9): 817–24. [PubMed] [Google Scholar]
[8] Patel V, Burns JK, Dhingra M, Tarver L, Kohrt BA, Lund C. Income inequality and depression: a systematic review and meta-analysis of the association and a scoping review of mechanisms. World Psychiatry2018;17(1):76–89. [PMC free article] [PubMed] [Google Scholar]
[9] Ho KY, Li WHC, Chan SSC. The effect of poverty and income disparity on the psychological well-being of Hong Kong children. Public Health Nurs2015;32(3):212–21. [PubMed] [Google Scholar]
[10] Pabayo R, Kawachi I, Gilman SE. Income inequality among American states and the incidence of major depression. J Epidemiol Community Health2014;68(2):110–5. [PMC free article] [PubMed] [Google Scholar]
[11] Pabayo R, Dunn EC, Gilman SE, Kawachi I, Molnar BE. Income inequality within urban settings and depressive symptoms among adolescents. J Epidemiol Community Health2016;70(10):997–1003. [PMC free article] [PubMed] [Google Scholar]
[12] Makaroun LK, Brown RT, Diaz-Ramirez LG, Ahalt C, Boscardin WJ, Lang-Brown S, et al. Wealth-associated disparities in death and disability in the United States and England. JAMA Intern Med2017;177(12):1745. [PMC free article] [PubMed] [Google Scholar]
[13] Wang Z, Du W, Pang L, Zhang L, Chen G, Zheng X. Wealth inequality and mental disability among the Chinese population: a population based study. Int J Environ Res Public Health2015;12(10):13104–17. [PMC free article] [PubMed] [Google Scholar]
[14] Dijkstra-Kersten SM, Biesheuvel-Leliefeld KE, van der Wouden JC, Penninx BW, van Marwijk HW. Associations of financial strain and income with depressive and anxiety disorders. J Epidemiol Community Health2015;69(7):660–5. [PubMed] [Google Scholar]
[15] Johnson W, Krueger RF. How money buys happiness: genetic and environmental processes linking finances and life satisfaction. J Pers Soc Psychol2006;90(4):680–91. [PubMed] [Google Scholar]
[16] Scholten S, Velten J, Neher T, Margraf J. Wealth, justice and freedom: objective and subjective measures predicting poor mental health in a study across eight countries. SSM Popul Health2017;3:639–48. [PMC free article] [PubMed] [Google Scholar]
[17] Yiengprugsawan V, Lazzarino AI, Steptoe A, Seubsman S, Sleigh AC. Psychosocial job characteristics, wealth, and culture: differential effects on mental health in the UK and Thailand. Global Health2015;11(1):31. [PMC free article] [PubMed] [Google Scholar]
[18] Lê-Scherban F, Brenner AB, Schoeni RF. Childhood family wealth and mental health in a national cohort of young adults. SSM Popul Health2016;2:798–806. [PMC free article] [PubMed] [Google Scholar]
[19] National Center for Health Statistics (NCHS) Centers for Disease Control, Prevention (CDC). DHANES_Response required: priority medium; Mode WebForm; Topic NHANES Depress screener eligibility. National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention (CDC); 2019. [Google Scholar]
[20] Kroenke K, Spitzer RL, Williams JBW. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med2001;16(9):606–13. [PMC free article] [PubMed] [Google Scholar]
[21] Hansson M, Chotai J, Nordstöm A, Bodlund O. Comparison of two self-rating scales to detect depression: HADS and PHQ-9. Br J Gen Pract2009;59(566): e283–8. [PMC free article] [PubMed] [Google Scholar]
[22] StataCorp.Stata statistical software. College Station, TX: StataCorp LLC; 2018. [Google Scholar]
[23] Pool LR, Needham BL, Burgard SA, Elliott MR, de Leon CFM. Negative wealth shock and short-term changes in depressive symptoms and medication adherence among late middle-aged adults. J Epidemiol Community Health2017;71(8):758–63. [PubMed] [Google Scholar]
[24] Frank C, Davis CG, Elgar FJ. Financial strain, social capital, and perceived health during economic recession: a longitudinal survey in rural Canada. Anxiety Stress Coping2014;27(4):422–38. [PubMed] [Google Scholar]
[25] Wilkinson LR. Financial strain and mental health among older adults during the great recession. J Gerontol B Psychol Sci Soc Sci2016;71(4):745–54. [PubMed] [Google Scholar]
[26] Adams DR, Meyers SA, Beidas RS. The relationship between financial strain, perceived stress, psychological symptoms, and academic and social integration in undergraduate students. J Am Coll Health2016;64(5):362–70. [PMC free article] [PubMed] [Google Scholar]
[27] Kendler KS, Karkowski LM, Prescott CA. Causal relationship between stressful life events and the onset of major depression. Am J Psychiatry1999;156(6): 837–41. [PubMed] [Google Scholar]
[28] Kessler RC. The effects of stressful life events on depression. Annu Rev Psychol1997;48(1):191–214. [PubMed] [Google Scholar]
[29] Brilman EI, Ormel J. Life events, difficulties and onset of depressive episodes in later life. Psychol Med2001;31(5):859–69. [PubMed] [Google Scholar]
[30] The Federal Reserve Board of Governors. Dealing with unexpected expenses. https://www.federalreserve.gov/publications/2018-economic-well-being-of-us-households-in-2017-dealing-with-unexpected-expenses.htm. [Accessed 8 November 2019].
