Interpersonal experiences with perceived racism cause emotional and psychological harm resulting in trauma-related symptoms such as depression, disordered sleep, hyper vigilance, isolation, and overall poor health [1
]. Additionally, systemic racism such as disparate sentencing laws based on race/ethnicity, housing discrimination, discrimination in health care, and disparities in school funding have also been found to contribute to poor health [5
]. Taken together, discrimination in a variety of settings, whether interpersonal or experienced through the systems that disproportionately affect the functions of everyday living, is likely at the root of racial and ethnic disparities in health and their social determinants, such as household food insecurity (HFI). For example, in 2018, 21.2% of African American or Black households and 16.2% of Hispanic households reported food insecurity, compared to 8.1% of white households [6
]. Similarly, in a recent nationwide study, 15% of Black adults and 11% of Hispanic adults reported being in fair or poor physical health, compared to 9% of white adults [7
]. Black individuals are also more likely than non-Latinx whites to report depressive symptoms, though reports comparing differences between Latinx and non-Latinx whites are mixed [8
While there is ample evidence of persistent racial and ethnic health disparities, there are methodological weaknesses in the conventional public health approach—namely, its emphasis on individual mechanisms rather than the subtle and systemic ways racism shapes access to opportunities in education, employment, housing and neighborhood resources and affects individual and collective physical, mental, and spiritual health [9
]. Critical Race Theory calls for attention to equity and critical self-reflection from investigators and public health practitioners; this approach demands that we take a hard look at previous methods and improve current methods to address issues of systemic racism [10
]. Using the lens of Critical Race Theory and given (1) the paucity of studies that investigate health impacts of discrimination in specific settings among mothers of young children and (2) the limited research that simultaneously investigates the impact of discrimination among multiple racial/ethnic groups [11
], the current study sought to examine the associations between race and ethnicity-based discrimination and multiple health outcomes. More specifically, we hypothesized that among mothers of children under age 4, lifetime experiences of discrimination (EODs) in a variety of settings, such as by police and courts, in public, at work and in school [12
], would be associated with HFI, caregiver depressive symptoms, and fair/poor physical health, and that there would be differential associations by race and ethnicity. Understanding that EODs are a form of adversity linked to trauma-related symptoms, we also assessed EOD relations to these outcomes while taking into account exposure to adverse childhood experiences (ACEs), which include experiences with abuse, neglect, and household adversity such as witnessing violence or having a parent in prison. ACEs have long been associated with major negative health outcomes, such as cardiovascular disease, diabetes, and depression [13
], and have also shown significant association with household food insecurity [16
]. We hypothesized that EODs would be associated with poorer health outcomes even after accounting for past exposure to ACEs. Finally, given evidence that nutrition assistance programs help reduce household food insecurity and related poor health [19
], we hypothesized that public assistance participation may attenuate or eliminate these associations.
3.1. Participant Characteristics and Experiences of Discrimination
Mean caregiver age was 27.4 years; mean child age was 19.7 months (Table 1
). Over half of the mothers (57.3%) self-identified as Latinx (Hispanic, Latino(a), or Spanish), 33.2 percent identified as Black non-Latinx, 7.9 percent identified as white non-Latinx, and 3.4 percent identified as another race/ethnicity or more than one race/ethnicity. The majority of mothers (85.8%) were born in the United States or Puerto Rico. Immigrant mothers were primarily from the Caribbean (7.1%), Mexico, and Central America (3.9%). Childhood adversity was prevalent with 35.6% of mothers reporting 1–3 ACEs and 12.0% of mothers reporting ≥4 ACEs. Eighty-two percent of mothers participated in public insurance (Medicare, Medicaid, military healthcare benefits) and most mothers reported receiving some form of public assistance. For example, 76.3% participated in SNAP, and 71.3% participated in WIC.
The most prevalent settings for discrimination experiences were in public or on the street (27.2%), stores and restaurants (20.1%), and at work (20.1%) (Table 2
). Rates of experiences of discrimination were less than five percent in settings such as housing, medical care, and banking. Reports of experiences of discrimination were also low in applying for public assistance (6.9%, n = 94). EODs varied significantly by race/ethnicity across all settings except housing, medical care, and applying for public assistance. Apart from medical care and applying for public assistance benefits, rates of discrimination were highest among non-Latinx Black mothers.
3.2. Food Insecurity, Depressive Symptoms and Poor Health
As stated above, after collapsing some EOD items into broader categories, only the most common EOD items (prevalent among at least 10% of mothers: police or in courts, public places, hiring and workplace, school) were modeled in regression analyses. EODs were associated with HFI and health outcomes (Table 3
). Adjusting for maternal age, race/ethnicity, educational attainment, years lived in the United States, insurance coverage, employment, marital status, and household size (Model 1), mothers who reported EODs from police or in courts had more than double the odds of HFI compared to mothers with no EODs from police or in courts (AOR = 2.54, 95% CI: 1.82–3.53). With further adjustment for ACEs in model 2, this association was attenuated, but remained statistically significant (Model 2 AOR = 2.06, 95% CI: 1.45–2.91). Subsequent adjustment for receipt of public assistance benefits in model 3 did not significantly change the strength of association (AOR = 2.04, 95% CI: 1.44–2.89). EODs in hiring and workplaces were similarly associated with higher odds of HFI (Model 3 AOR = 1.57, 95% CI: 1.18–2.11), as were EODs in school (Model 3 AOR = 1.44, 95%CI: 1.06–1.96). EODs in public places was associated with higher odds of HFI. However, once adjusted for ACEs and public assistance participation, this association was no longer statistically significant.
EODs in all settings were associated with elevated odds of depressive symptoms, ranging from AOR = 1.65 (95% CI: 1.16–2.33) for EODs in public places to AOR = 1.96 (95% CI: 1.30–2.97) for EODs in school. However, for EODs in all settings these associations were attenuated and no longer statistically significant after further adjustment for ACEs and public assistance benefits.
EODs from police or in courts, in public places, and in hiring and workplaces were all associated with higher odds of poor physical health status (Model 1: AOR = 1.47 [95% CI: 1.02–2.12], AOR = 1.59 [95% CI: 1.22–2.08], and AOR = 1.36 [95% CI: 1.02–1.83], respectively). The association of EODs in public places with poor physical health status was somewhat attenuated but remained statistically significant after adjusting for ACEs in model 2, (AOR = 1.36, 95% CI: 1.01–1.83). Further adjustment for receipt of public benefits in model 3 did not alter the association (AOR = 1.37, 95% CI: 1.02–1.84). Associations of physical health status with EODs from police or in courts and in hiring and workplaces were more strongly attenuated and no longer statistically significant after further adjustment for ACEs. In contrast, EODs in school was not associated with poor physical health in model 1 (AOR = 1.04, 95% CI: 0.99–1.09), but was associated with higher odds of poor physical health after further adjustment for ACEs in model 2 (AOR = 1.95, 95% CI: 1.28–2.97). This association was somewhat attenuated after additional adjustment for receipt of public assistance benefits in model 3 (AOR = 1.50, 95% CI: 1.09–2.08).
3.3. Outcomes by Race/Ethnicity and Experiences of Discrimination
In fully adjusted models stratified by mothers’ race/ethnicity, associations between EODs and the outcomes were largest among Latinx mothers (Table 4
). EODs in at least one setting was associated with every outcome among Latinx mothers, while among Black mothers only HFI was associated with EODs in some settings and among white mothers none of the outcomes were statistically significantly associated with EODs in any setting. Among Latinx mothers, those who reported EODs had higher odds of HFI compared to those who reported no EODs in the same setting for EODs from police or in courts (AOR = 2.14, 95% CI: 1.33–3.45), in hiring and workplaces (AOR = 1.76, 95% CI: 1.21–2.56), and in school (AOR = 1.58, 95% CI: 1.05–2.36), but not in public places (AOR = 1.38, 95% CI: 0.97–1.98). Among Black mothers, odds of HFI were also elevated for EODs from police or in courts (AOR = 2.00, 95% CI: 1.32–2.96) and in hiring and the workplace (AOR = 1.46, 95% CI: 1.05–2.04), but not for EODs in public places or in school. Although not statistically significant, the estimate for the association between EODs from police or in courts and HFI was similar among non-Latinx white mothers as among mothers from the other groups (AOR = 1.83, 95% CI: 0.81–4.12).
EODs in school were associated with almost 80 percent higher odds of depressive symptoms among Latinx mothers (AOR = 1.79, 95% CI: 1.12–2.86). EODs in school were not associated with depressive symptoms among Black or white mothers, and EODs in other settings were not associated with depressive symptoms in any group. Among Latinx mothers only, poor physical health status was also associated with EODs in public places (AOR = 1.63, 95% CI: 1.12–2.38) and EODs in school (AOR = 1.96, 95% CI: 1.29–2.99).
Lifetime experiences of racial and ethnic discrimination are associated with HFI, depressive symptoms, and fair/poor physical health for mothers of young children—in some cases, independently of ACEs. These results corroborate results found by Burke et al. [34
], where lifetime discrimination in workplaces and schools was associated with the increased severity of HFI among a sample of 154 African American households. Our results were obtained with a significantly larger and more diverse sample, allowing us to demonstrate that the association between EODs and negative health outcomes persists despite participation in public benefit programs. Higher odds of poor health among mothers who experienced discrimination in public as well as higher odds of HFI among mothers who experienced discrimination from the police or courts and in the workplace were attenuated by adjustment for ACEs and public assistance but remained statistically significant. Our results also demonstrated that discrimination experienced in different settings has varying relationships with physical and mental health. For example, while discrimination experienced in public places was not associated with HFI, it was associated with poor physical health.
While in most cases adjustment for ACEs attenuated associations of EODs with the outcomes, suggesting the importance of accounting for other adversities when investigating the health impacts of EODs, the association between discrimination experienced in school and HFI and with physical health was strengthened by adjustment for ACEs. A plausible explanation for this particular change is that EODs in school may overlap with experiences of being bullied. Nearly 1 in 10 individuals with high exposures to ACEs also experience childhood bullying [35
]. Victims of bullying have increased risk of poor health, dropping out of school, unstable employment, and living in poverty, independent of hardships and family dysfunction in childhood [36
]. Thus, adjusting for ACEs may have captured the independent impacts of bullying rather than discrimination per se, which may explain the negative confounding phenomenon demonstrated in the results.
Although Black non-Latinx mothers reported the highest rates of discrimination, increases in odds of poor outcomes other than HFI associated with discrimination were statistically significant only for Latinx mothers. This may be due to the higher number of Latinx mothers in our sample, leading to greater ability to detect associations in this group. Or, it may be due to unmeasured factors in how Latinx mothers experience and react to discrimination which may be different from non-Latinx Black or white mothers. It should also be noted that the current study’s use of perceived experiences of discrimination, though important, does not fully capture total exposure to discrimination, which can go far beyond perceived experiences [37
]. Experiences of institutional discrimination, such as residential segregation, differential access to societal goods and resources, and disparate sentencing laws, which were not accounted for in the study analyses, also adversely affect health and can compound the effects of perceived interpersonal discrimination.
Similarly, the lack of association between discrimination and outcomes for non-Latinx white mothers may reflect the smaller sample and lower rates of reported discrimination in this group, as well as societal and structural factors that shield non-Latinx white mothers from effects of discrimination. For example, while the majority of adults living in the U.S. are white, poor white families are not residentially concentrated in ways that Black and Latinx families are, due to residential segregation and historically racist policies that exploit or exclude communities of color [39
]. Residential segregation leads to a “concentration of economic and social disadvantage and the absence of an infrastructure that promotes opportunity,” including: lower access to affordable, healthy foods, poorer housing quality, and increased exposure to environmental toxins, all of which are associated with worsened health and higher rates of premature death [5
There are many plausible pathways between discrimination and HFI and health. Discrimination by police and courts has been linked to higher incarceration rates, which negatively impacts income, food security, and health [43
]. Workplace discrimination based on race, ethnicity and gender can affect wages, job security, and ability to be promoted, which in turn affects income and mental health [44
]. Finally, discrimination in schools is associated with differences in disciplinary actions that affect school performance which in turn affects mental health and income [45
This study has a cross-sectional design, hence a temporal, causal link between discrimination and poor health cannot be made. Study results are also limited in that the mothers are from a single pediatric hospital; there was no regional variation to situate exposure to discrimination, health, and HFI within larger societal forces, such as housing segregation and disparate school funding. Also, health outcomes were self-reported rather than diagnosed by a health care provider, although the measures have been validated for a wide variety of settings. Recruiting participants through a hospital emergency department may introduce selection bias toward participants with poorer health overall, although in our study the mothers were at the emergency department to seek care for their children, not themselves, and the majority of mothers reported excellent or good physical health. Although fewer than 5 percent of mothers in the sample reported discrimination while getting medical care, it is possible that prior experiences of discrimination in health care settings might dissuade mothers from seeking care at the emergency department, leading to an underrepresentation in our sample of mothers who had experienced discrimination in this setting. Finally, due to the nature of our sample, our paper does not take into account the intersectional nature of gender and race/ethnicity to account for known ways in which the confluence of gender, racial and ethnic discrimination impact poor health and food insecurity [46
Despite its limitations, this study contributes to the public health literature in several ways. First, it investigates the relationship between discrimination, HFI and health across multiple racial/ethnic groups using a large sample demonstrating that, while discrimination is harmful for everyone, it may be particularly harmful to Black and Latinx families. It also accounts for compounded adversities during childhood such as abuse, neglect, and household dysfunction captured by ACEs. Our study both answers the call for greater use of and adds to the evidence base for Critical Race Theory, which integrates self-reflection on methods and research questions, and thorough attention to dynamics of racism and poor health in order to get to the roots of poor health and health disparities [10
]. Our study answers the call by bringing focused attention to discrimination in a variety of settings and assessing their impact on food insecurity and health among women and their young children. It adds to the evidence base by empirically demonstrating that the relationship between interpersonal discrimination and worsened health persists despite receipt of public benefits. In turn, this illuminates the mechanism(s) by which larger societal forces of racial and ethnic discrimination can contribute to poor health and adversity.
Our study’s finding that public assistance did not mitigate the relationship between discrimination and poor health, sheds light on how public benefits meant to address HFI and health may fall short. This may be for a variety of reasons rarely considered by researchers such as social position, differential treatment by the police and courts, in schools and workplaces, as well as potentially experiencing discrimination while shopping for food [51
]. We also join others to frame these experiences with racial/ethnic discrimination as a form of violence exposure that is fundamentally associated with HFI, poor physical health and depressive symptoms [52
]. This study highlights the severity of the associations between discrimination and health outcomes, above and beyond other trauma (such as ACEs) and in spite of programs intended to improve health, such as health insurance and receipt of public assistance. Future research should take into account the intersectional nature of discrimination to ensure that public health programs attend to issues of gender and gender expression while addressing racial and ethnic disparities.