Experiences of Racial Discrimination and Periconceptional Diet Quality

Abstract

Background

Racism is a key determinant of perinatal health disparities. Poor diet may contribute to this effect, but research on racism and dietary patterns is limited.

Objective

We aimed to describe the relation between experiences of racial discrimination and adherence to the 2015‒2020 Dietary Guidelines for Americans.

Methods

We used data from a prospective pregnancy cohort study conducted at 8 United States medical centers (2010‒2013). At 6‒13 weeks of gestation, 10,038 nulliparous people with singleton pregnancies were enrolled. Participants completed a Block food frequency questionnaire, assessing usual diet in the 3 mo around conception, and the Krieger Experiences of Discrimination Scale, assessing the number of situational domains (e.g., at school and on the street) in which participants ever experienced racial discrimination. Alignment of dietary intake with the 2015–2020 Dietary Guidelines for Americans was assessed using the Healthy Eating Index (HEI)-2015.

Results

The study showed that 49%, 44%, 35%, and 17% of the Asian, Black, Hispanic, and White participants reported experiences of racial discrimination in any domain. Most participants experienced discrimination in 1 or 2 situational domains. There were no meaningful differences in HEI-2015 total or component scores in any racial or ethnic group according to count of self-reported domains in which individuals experienced discrimination. For example, mean total scores were 57‒59 among Black, 61‒66 among White, 61‒63 among Hispanic, and 66‒69 among Asian participants across the count of racial discrimination domains.

Conclusions

This null association stresses the importance of going beyond interpersonal racial discrimination to consider the institutions, systems, and practices affecting racialized people to eliminate persistent inequalities in diet and perinatal health.

Introduction

Black birthing people and infants in the United States disproportionately bear the burden of adverse pregnancy and birth outcomes compared with most other racial and ethnic groups [1]. For decades, the absolute incidence of poor perinatal health outcomes has declined; however, the rate of maternal and infant mortality among Black individuals has remained 2‒4 times that of their Hispanic, Asian, or White counterparts [2,3]. There is growing recognition that racism, and particularly racial discrimination, is a key driver of disparities in perinatal health [4,5]. Racial discrimination, defined as unfavorable or unfair treatment based on race or personal characteristics associated with race (such as hair texture, skin color, or certain facial features) [6], affects maternal and child health outcomes through numerous pathways, such as socioeconomic disadvantage, access and quality of prenatal care, stress, and toxic exposures [[7], [8], [9], [10]]. An additional pathway may be through dietary patterns.

A healthy diet before and during pregnancy has been associated with fewer adverse perinatal health outcomes [11]. Racial discrimination may link with adverse perinatal outcomes through 2 dietary behavior mechanisms [10]. First, racial discrimination can limit an individual’s or family’s material resources, such as access to income, wealth, living or thriving wage employment, and supportive neighborhood assets (e.g. supermarkets), which contribute to poor dietary intake [12]. Second, racial discrimination contributes to psychosocial stress that also can affect eating behaviors, such as contributing to overeating and poor dietary quality [13]. The inextricable connection between unhealthy environments and stress means that these 2 pathways are not mutually exclusive [13]. Nevertheless, there is limited research into the association between racism and adherence to dietary recommendations. A recent systematic review [14] found mixed results in the 6 studies evaluating self-reported racial discrimination and diet quality [[15], [16], [17], [18], [19], [20]], but only half used tools that had undergone some form of validation [[15], [16], [17]], and none were conducted among pregnant individuals.

Our objective was to advance understanding of the relationship between interpersonal racial discrimination and dietary patterns during pregnancy by describing the association between self-reported experiences of racial discrimination and adherence to the Dietary Guidelines for Americans. In addition, we stratified these associations by education as a proxy for economic marginalization, recognizing that the intersection between racism and economic marginalization affect dietary intake [21].

Materials and Methods

We used data from the Nulliparous Pregnancy Outcomes Study: monitoring mothers-to-be (nuMoM2b), a pregnancy cohort study of 10,038 pregnant individuals receiving care at one of the 8 United States medical centers from 2010 to 2013. Details of the study can be found elsewhere [22,23]. Eligibility criteria included singleton pregnancy, 6 to 13 completed weeks of gestation at enrollment, and no previous pregnancy lasting ≥20 weeks of gestation. Trained and credentialed study personnel and a common protocol and manual of operations were used at all sites. At enrollment (6‒13 completed weeks’ gestation), usual periconceptional dietary intake was ascertained, along with data on demographics, medical history, and psychosocial and behavioral factors. At the second-trimester study visit, data on experiences of racial discrimination were collected. Every site’s local institutional review board approved the study, and all participants gave written, informed consent.

Participants were asked whether they are of Hispanic or Latino origin or descent and which group or groups describe their racial background: White; Black, African American, or African Descent; American Indian or Alaska Native; Native Hawaiian or Other Pacific Islander; Asian Indian; Other Asian; or Other. We classified individuals into nonmutually exclusive groups: Black or not Black, White or not White, Hispanic or not Hispanic (regardless of self-reported race), and Asian or not Asian. There were too few American Indians, Alaska Natives, Native Hawaiians, or Other Pacific Islanders to separate this group, but most self-reported their ethnicity as Hispanic and, thus, were included in the Hispanic group.

Self-reported experiences of racial discrimination were collected using the “situation version” of the 9-item Krieger Experiences of Discrimination Scale [24]. This tool has been demonstrated to have acceptable validity and reliability in Black, Latino, and White adults [24]. Participants were asked, “Have you ever experienced discrimination, been prevented from doing something, or been hassled or made to feel inferior in any of the following situations because of your race, ethnicity, or color?” Yes or no responses were given for 9 possible situational domains: at school; getting hired or getting a job; at work; getting housing; getting medical care; getting service in a store or restaurant; getting credit, bank loans, or a mortgage; on the street or in a public setting; from the police; or in the courts. Unlike later versions of this tool, the version in our study did not include timing or frequency of racial discrimination in these situational domains. The number of domains reported were summed for a measure of total situational domains that ranged from 0 to 9. We categorized participants into 0, 1 to 2, 3, or 4 or more situational domains based on the distribution of the scores in our cohort and previous research [24,25].

At enrollment, participants completed a self-administered modified Block 2005 FFQ (available in English and Spanish) on paper to assess usual dietary intake in the 3 mo around conception. For easier recall of the 3 mo around conception, a trained study staff member labeled each woman’s FFQ with the months of interest: the month before their last menstrual period, the month of their last period, and the month following their last period. The FFQ assessed usual frequency and portion size of 119 food and beverage items. The FFQ was modified to focus on the period 3 mo around the start of pregnancy, to include foods more common in the diets of Spanish-speaking populations, and to add fish items for better capturing of omega-3 intake. The FFQ also had 33 additional questions to ascertain more detail regarding fat (e.g., low-fat or fat-free), sodium (i.e., low-sodium), and carbohydrate (i.e., low-carb or low-sugar) contents of certain foods, in addition to types of fats used in cooking and ready-to-eat cereal consumed. Study personnel checked all pages of the FFQ for completeness.

The FFQ has acceptable validity (most correlations 0.5‒0.6 for nutrients in comparison with 4-d food records), including in racially diverse samples [[26], [27], [28], [29]]. Block Dietary Data Systems performed scanning, nutrient and food group mapping, and a summary analysis of the FFQ [30]. The food items were disaggregated into their component parts and linked to the nutrient database that was developed based on the USDA Food and Nutrient Database for Dietary Studies [31] and the Food Patterns Equivalents Database [31].

We evaluated the extent to which dietary intake aligned with the 2015–2020 Dietary Guidelines for Americans [32] using the Healthy Eating Index (HEI)-2015 [33]. The HEI-2015 is appropriate for use in pregnancy [32]. The HEI-2015 includes 9 “adequacy” components (higher intake is advised): total fruits, whole fruits, total vegetables, greens and beans, whole grains, dairy, total protein foods, seafood and plant proteins, and fatty acids. It also includes 4 “moderation” components (limited intake is advised): refined grains, sodium, added sugars, and saturated fats. Components have maximum scores of 5 or 10. The criteria for component scores are summarized in Supplemental Table 1. We calculated person-level HEI-2015 total and component scores. The total score reflects overall diet quality, whereas the component scores provide information on the food groups and components for which intake was closer to or farther from the guidance [33].

At enrollment, participants self-reported their highest level of education, which we classified as high school or less (less than high school or high school graduate or equivalent), some college (some college credit, but no degree, or associate/technical degree), or college graduate (bachelor degree or higher). Other information self-reported at the first visit included marital status, smoking before pregnancy, medical insurance, employment status, whether their pregnancy was planned, and country of their birth and their parents’ births. Prepregnancy BMI was calculated as self-reported prepregnancy weight (reported at the first visit) (kg) divided by measured height (m) squared. We classified BMI (in kg/m²) based on thresholds: <18.5, 18.5‒24.9, 25‒29.9, or ≥30. The United States census-based neighborhood and built environment variables, which were based on census tract and block group details of the participants at enrollment, were neighborhood deprivation [34] and percentage of neighborhood with income below the poverty line. These 2 neighborhood variables were divided into 3 groups based on tertiles of their distributions.

Statistical analysis

Of the 10,038 participants, we excluded 15 individuals who did not identify their race or ethnicity. For the remaining 10,023 individuals, data were missing for dietary intake (n = 1779), experiences of discrimination (n = 642), or education or key covariates (n = 429). To address the missing data, we used multiple imputation by chained equations, imputing 10 data sets using all available variables [35].

We compared HEI-2015 total and component scores according to the number of situational domains in which the participant reported experiencing racial discrimination. Two effect modifiers were of interest: race or ethnicity and education. Our primary analyses were stratified by race or ethnicity and then according to the pregnant women's education among Black, White, and Hispanic participants. We could not perform the latter evaluation for the Asian race because of small numbers in some cells. We used maternal education as a proxy for economic marginalization because we lacked other relevant measures.

Our objective was to describe associations, as opposed to determining causal effects. Therefore, we did not fit multivariable regression models. Furthermore, we did not perform null-hypothesis significance testing based on guidance by the American Statistical Association [36]. Rather, we focused on meaningful differences, which the developers of the HEI-2015 suggest may be a 5-point to 6-point differences in total scores among groups (effect size of 0.5) [37]. We used the standard deviation from our data to determine mean differences based on this effect size [38]. Standard deviations were all 10 or greater, leading to a 6-point difference in means, considered as meaningful. We extrapolated this method for determining meaningful differences in component scores based on standard deviations from our data as 1.3-point or 0.8-point differences in 10-point or 5-point components, respectively. To provide a visual representation of dietary pattern [39], we created radar plots, with each HEI-2015 component score depicted as a mean of the percentage of maximum points for each discrimination subgroup on its respective axis. Means were compared for the total and component scores after evaluating skewness of the distributions.

Results

Almost half of the Black (44%) or Asian (49%) individuals reported experiencing racial discrimination, followed by 35% of the Hispanic and 17% of the White participant (Table 1). The top 3 most common situational domains in which Black, Hispanic, and Asian participants reported ever experiencing discrimination were at school, while receiving service at a store or restaurant, and on the street or in public. Most people who reported racial discrimination experienced it in 1 or 2 situational domains.

TABLE 1.

Descriptive characteristics of the cohort by race or ethnicity and the number of situational domains in which participants reported ever experiencing racial discrimination: Nulliparous Pregnancy Outcomes Study: monitoring mothers-to-be (2010‒2013).

	Black				White				Hispanic				Asian
	Reported number of situational domains experiencing racial discrimination				Reported number of situational domains experiencing racial discrimination				Reported number of situational domains experiencing racial discrimination				Reported number of situational domains experiencing racial discrimination
	0, %	1 or 2, %	3, %	4 or more, %	0, %	1 or 2, %	3, %	4 or more, %	0, %	1 or 2, %	3, %	4 or more, %	0, %	1 or 2, %	3, %	4 or more, %
n	985	450	147	179	5866	958	138	91	1110	401	84	109	276	185	44	43
Age, y
<20	47	35	24	12	10	14	13	16	28	21	15	23	5	7	3	7
20–29	44	49	54	48	50	46	44	43	53	55	57	47	43	33	41	34
≥30	9	16	22	40	40	41	42	41	19	24	28	30	52	60	56	59
Education
High school or less	50	38	22	22	13	17	16	26	35	27	19	27	6	5	3	11
Some college	40	40	43	37	26	26	34	30	41	38	45	43	20	24	20	19
College degree	10	22	35	41	60	57	40	44	23	35	35	29	74	71	77	69
Marital status
Unmarried	89	81	77	70	25	36	42	62	62	56	46	68	19	16	20	27
Married	11	19	23	30	75	64	56	38	38	44	54	32	81	84	80	73
Health insurance
Private	29	37	48	50	81	73	67	55	37	45	58	39	80	84	88	78
Public	71	63	52	50	19	27	33	45	63	55	42	61	20	16	12	22
Smoking
Nonsmoker	75	73	70	70	84	78	73	68	87	86	79	77	92	92	88	97
Smoker	25	27	30	30	16	22	27	32	13	14	21	23	8	8	12	3
Prepregnancy BMI, kg/m2
<18.5	4	4	4	4	4	3	4	4	4	4	2	3	9	8	2	13
18.5‒24.9	43	33	35	31	59	56	46	45	53	47	51	45	69	68	68	50
25‒29.9	23	28	26	25	21	22	28	24	26	26	24	22	14	12	17	18
≥30	30	35	35	40	16	19	22	27	17	23	23	31	7	10	13	18
Currently employed
Yes	51	58	71	73	85	80	72	64	64	66	72	56	71	76	91	77
No	49	42	29	27	15	20	28	36	36	34	28	44	29	24	9	23
% of Neighborhood with income below poverty index
Lowest third	9	11	6	13	43	37	30	29	16	19	12	11	37	30	38	20
Middle third	56	28	28	30	36	36	38	30	30	31	40	25	36	45	43	43
Upper third	68	61	66	57	21	27	31	40	54	50	48	64	27	25	19	37
Area deprivation
Lowest third	7	12	12	16	40	38	37	23	28	31	29	20	54	47	60	52
Middle third	18	18	25	21	40	36	32	41	26	28	23	34	26	33	30	26
Upper third	75	69	63	62	20	26	31	36	45	41	47	47	20	19	10	22
Acculturation
United States born/parents United States born	87	84	83	84	84	72	68	61	23	21	22	22	11	8	6	12
United States born/parent(s) immigrants	7	8	11	11	9	15	16	19	34	37	42	35	30	30	56	33
Born outside United States, immigrated at younger than 18 y	2	5	5	3	4	8	8	11	26	27	19	30	25	35	15	22
Born outside United States, immigrated at 18 y or older	4	3	1	2	3	6	8	9	17	15	17	12	34	27	23	33

Black participants who reported discrimination across more situational domains were more likely than those reporting fewer situations to be older, college-educated, married, and employed, and have private insurance and live in a neighborhood with less deprivation (Table 2). Among White participants, situations of discrimination were most commonly reported by those who were unmarried, receiving public insurance, smokers, unemployed, born to immigrant parents, and living in a neighborhood with higher poverty and deprivation, and had a high school education or less and higher BMI. There were few clear patterns among Hispanic and Asian individuals by this racial discrimination measure.

TABLE 2.

Self-reported experiences of racial discrimination according to self-reported race or ethnicity: Nulliparous Pregnancy Outcomes Study: monitoring mothers-to-be (2010‒2013).

	Black (n = 1760), %	White (n = 7053), %	Hispanic (n = 1703), %	Asian (n = 547), %
Experienced discrimination1
Never	56	83	65	51
Ever	44	17	35	49
Situational domains in which participants reported ever experiencing racial discrimination
At school	18	6	14	27
Getting hired	11	3	8	8
At work	16	5	11	14
Getting housing	3	6	2	1
Receiving medical care	2	0	3	1
Service at store or restaurant	23	0	13	21
Getting credit, mortgage, or loan	2	3	2	1
On the street or in public	28	1	18	30
From police or the courts	10	8	6	4
No. of situations mentioned (possible range, 0–9)
0	56	83	65	51
1 or 2	26	14	24	34
3	8	2	5	8
≥4	10	1	6	7

¹Self-reported experiences of racial discrimination were collected using the previously validated “situation version” of the 9-item Kreiger Experiences of Discrimination Scale [24].

Mean (95% confidence interval) total HEI-2015 score were the lowest for Black [58 (57, 58)] and Hispanic [63 (62, 63)] participants, followed by the scores for White [66 (65, 66)] and Asian [69 (68, 70)] individuals. There were no meaningful differences in total HEI-2015 scores in any racial or ethnic group according to the number of situations in which they experienced racial discrimination (Table 3). For example, mean scores ranged from 57 to 59 among Black participants and 61 to 63 among Hispanic participants across subgroups with 0, 1‒2, 3, or 4 or more situational domains in which they experienced discrimination.

TABLE 3.

Total Healthy Eating Index-2015 score according to race or ethnicity and the number of situational domains in which participants reported ever experiencing racial discrimination: Nulliparous Pregnancy Outcomes Study: monitoring mothers-to-be (2010‒2013).

Self-reported race or ethnicity	Self-reported experiences of racial discrimination	n	Healthy Eating Index-2015, mean (95% CI)
Black
	0 situation	985	57 (56, 57)
	1 or 2 situations	450	58 (57, 59)
	3 situations	147	59 (57, 61)
	4 or more situations	179	59 (57, 61)
White
	0 situation	5866	66 (66, 66)
	1 or 2 situations	958	65 (64, 65)
	3 situations	138	64 (62, 66)
	4 or more situations	91	61 (58, 64)
Hispanic
	0 situation	1110	63 (62, 63)
	1 or 2 situations	401	63 (62, 63)
	3 situations	84	63 (61, 66)
	4 or more situations	109	61 (59, 63)
Asian
	0 situation	276	69 (68, 71)
	1 or 2 situations	185	68 (66, 70)
	3 situations	44	68 (65, 71)
	≥4 situations	43	66 (62, 70)

When stratified by education, higher mean HEI-2015 scores were observed among Black, White, and Hispanic participants with higher compared with those of participants with lower educational attainment (Table 4). However, there were no differences according to the number of domains in which people reported racial discrimination.

TABLE 4.

Total Healthy Eating Index (HEI)-2015 score according to race or ethnicity, education, and the number of situational domains in which participants reported ever experiencing racial discrimination: Nulliparous Pregnancy Outcomes Study: monitoring mothers-to-be (2010‒2013).

Self-reported race or ethnicity	No. of situational domains in which participants reported ever experiencing racial discrimination	High school diploma or less		Some college or associate degree		Bachelor degree or higher
		n	Mean HEI-2015 (95% CI)	n	Mean HEI-2015 (95% CI)	n	Mean HEI-2015 (95% CI)
Black
	0 situation	495	54 (53, 55)	390	57 (56, 59)	99	66 (63, 68)
	1 or 2 situations	171	54 (52, 56)	177	57 (56, 59)	101	65 (63, 68)
	3 situations	33	52 (48, 56)	64	58 (55, 61)	51	66 (62, 69)
	≥4 situations	38	53 (49, 57)	57	58 (55, 61)	73	64 (61, 67)
White
	0 situation	766	56 (55, 57)	1541	63 (62, 63)	3557	70 (69, 70)
	1 or 2 situations	161	55 (53, 56)	247	62 (60, 63)	548	69 (68, 70)
	3 situations	22	52 (47, 57)	48	61 (58, 64)	69	69 (67, 72)
	≥4 situations	23	52 (47, 56)	27	58 (53, 64)	40	68 (64, 72)
Hispanic
	0 situation	394	59 (58, 61)	456	62 (61, 63)	259	69 (67, 70)
	1 or 2 situations	107	60 (58, 62)	150	62 (60, 64)	143	66 (64, 68)
	3 situations	16	56 (49, 63)	37	62 (59, 66)	29	68 (65, 72)
	≥4 situations	29	59 (55, 63)	45	60 (56, 64)	31	65 (60, 70)

In general, radar plots did not show clear differences in the alignment of intake with guidance for the HEI-2015 components comparing participants reporting racial discrimination across 0, 1‒2, 3, or 4 or more situational domains for Black (Figure 1), White (Figure 2), Hispanic (Figure 3), or Asian (Figure 4) individuals. The only meaningful differences in component scores by counts of discrimination domains were for refined grains (Black and Asian) and whole grains (White) (Supplemental Table 2). Among Hispanics, there were no meaningful differences.

FIGURE 1.

HEI-2015 component scores as a percentage of the maximum among Black participants (n = 1756) by the number of situational domains in which participants reported ever experiencing racial discrimination (0, blue; 1–2, red; 3, green; 4 or more, yellow). A perfect score is denoted by the blue-dotted line. Each HEI-2015 component score depicted as a mean of the percentage of maximum points for each discrimination subgroup on its respective axis. HEI, Healthy Eating Index.

FIGURE 2.

HEI-2015 component scores as a percentage of the maximum score among White participants (n = 7047) by the number of situational domains in which participants reported ever experiencing racial discrimination (0, navy blue; 1–2, gray; 3, orange; 4 or more, pink). A perfect score is denoted by the blue-dotted line. Each HEI-2015 component score depicted as a mean of the percentage of maximum points for each discrimination subgroup on its respective axis. HEI, Healthy Eating Index.

FIGURE 3.

HEI-2015 component scores as a percentage of the maximum score among Hispanic participants (n = 1696) by the number of situational domains in which participants reported ever experiencing racial discrimination (0, navy blue; 1–2, gray; 3, orange; 4 or more, pink). A perfect score is denoted by the blue-dotted line. Each HEI-2015 component score depicted as a mean of the percentage of maximum points for each discrimination subgroup on its respective axis. HEI, Healthy Eating Index.

FIGURE 4.

HEI-2015 component scores as a percentage of the maximum score among Asian participants (n = 547) by the number of situational domains in which participants reported ever experiencing racial discrimination (0, navy blue; 1–2, gray; 3, orange; 4 or more, pink). A perfect score is denoted by the blue-dotted line. Each HEI-2015 component score depicted as a mean of the percentage of maximum points for each discrimination subgroup on its respective axis. HEI, Healthy Eating Index.

Discussion

In this large, United States pregnancy cohort, individuals who identified as Black, Hispanic, or Asian commonly reported exposure to racial discrimination, particularly at school, while receiving service at a store or restaurant, and on the street or in public. However, the number of situational domains in which participants reported experiencing racial discrimination was not associated with overall adherence to national dietary guidelines in any racial/ethnic group. Stratification by education did not alter these conclusions.

There is a paucity of research relating racial discrimination to diet quality [14], with no previous work evaluating adherence to national dietary recommendations or studying populations of Hispanic or pregnant individuals. Only 3 previous studies, 2 in Black [15,17] and 1 in South Asian nonpregnant adults [16], evaluated this association using tools to assess racial discrimination and dietary intake that have been evaluated for validity. In South Asian participants, researchers reported an association between unfair treatment and higher intake of added sugars and sweets [16]. In a lifestyle modification trial within a community-based primary care sample of low-income Black hypertensive patients, investigators calculated a diet score from the Rapid Eating and Activity Assessment for Patients 3-point Likert scales for the frequency of eating 27 dietary items [17]. They found that previous exposure to discrimination was associated with a lower healthy diet score at baseline but a greater improvement in this score by 12 mo follow-up [17]. By contrast, 2 studies reported that unfair treatment over the lifetime or in the past year was not related to fruit and vegetable intake [15,16]. The remaining studies evaluating exposure to racial discrimination and diet quality used nonstandard dietary assessment tools, and their results were also mixed [[18], [19], [20]]. More robust evidence points to food insecurity as a consequence of racial discrimination in nonpregnant and pregnant individuals [[40], [41], [42], [43], [44]]. Food insecurity is associated with poor diet quality, such as inadequate intakes of fruits, vegetables, grains, meat, folate, iron, and calcium [45]. Unfortunately, we lacked a measure of food security in our sample.

Our null findings between experiences of racial discrimination and diet quality may be related to our measure of discrimination. The tool we used asked participants to indicate whether they ever experienced racial discrimination in 9 situational domains [24]. Although this tool has been used previously to link the count of the domains in which discrimination was experienced with outcomes such as hypertension [25], antenatal depressive symptoms [46], and fetal weight [43], it does not allow a more detailed interrogation of discrimination exposure. Quantifying the frequency, duration, and severity of unfair treatment in each domain, or the age at which it was experienced, may be more relevant to dietary behaviors.

More research should explore the pathways linking racial discrimination, diet, and adverse perinatal outcomes. Increased experiences with racial discrimination may contribute to a person’s socioeconomic position and limit economic mobility, which can restrict opportunities for accessing healthy food and contribute to psychosocial stress [13,47]. However, it is common that in non-Hispanic Black households, income is not related to perceived racial discrimination, with both low-income and high-income Black people reporting the highest levels of discrimination [48]. This is in contrast to non-Hispanic White households, in which household income is inversely associated with perceived racial discrimination [3]. In addition, and consistent with our findings, educational attainment in Blacks has been correlated with more, not less, perceived racial discrimination [48,49].

Regardless of income, racialized communities in the United States frequently experience restricted food environments [50,51], where availability of healthy foods from high-quality food outlets is limited, and relatively inexpensive energy-dense foods from corner and convenience stories are plentiful [12]. Unrelated to the food environment are other barriers to healthy diets, including social factors such as violence, crime, and poverty [12] and policies limiting access to affordable housing, child care, paid family leave, and livable wages [52,53]. Food and beverage marketing also influences dietary patterns, and African American communities are differentially exposed to media advertisements for unhealthy and low-cost food and beverage products [54].

One of our study’s important findings was the high percentage of Asian participants reporting experiences of discrimination across any situational domain. Although our results and those from national studies indicate a higher prevalence of Asian populations’ adherence to the dietary guidelines [55], the increased attention on anti-Asian racism and xenophobia in the United States may have implications for dietary behaviors [56]. More work is needed to explore how these experiences may affect dietary behaviors across different racial or ethnic groups and the mechanisms for the associations.

Our study lacked data that would allow for a deeper understanding of discrimination and diet quality, such as diet and shopping behaviors and food insecurity. Our sample size limited our ability to group individuals based on less heterogeneous racial and ethnic identities. Understanding dietary patterns and experiences of discrimination among pregnant people in all United States racialized groups is needed. It is also important to consider other measures for economic marginalization beyond education. Results from this cohort of nulliparous individuals from urban settings may not be generalizable to the general pregnant population. We sacrificed some generalizability for the racial or ethnic and geographic diversity of this cohort.

Absolute intake estimates from FFQ data have been shown to be subject to a higher degree of systematic measurement error than other self-report methods [57,58]. Attenuation of associations and reduced precision commonly result from this error [59], but its effects cannot be certain without a formal quantitative bias analysis. However, our adjustment for energy intake through the density approach helps to reduce the bias [60]. We used a threshold to identify meaningful differences in mean HEI-2015 total and component scores that was suggested by the creators of this index [37]. This approach may be viewed as too conservative, but we are not aware of other proposed thresholds in the literature. We did not seek to quantify causal effects of racial discrimination on diet quality. Rather, we described the distribution of the HEI-2015 scores according to situational domains of racial discrimination in the societal context of systemic racism. The null relation we observed can serve to motivate causal hypotheses to be tested in future research [61].

To reduce inequities in adherence to federal dietary guidance, studying the exposures and experiences of racialized communities in the United States is essential. Understanding the effect of racism on perinatal health, such as the complex relationships among structural, interpersonal, and internalized racism [62], has implications for developing effective programmatic and policy interventions to improve maternal and child health disparities in the United States.

Data availability

The Nulliparous Pregnancy Outcomes Study: monitoring mothers-to-be data described in the manuscript and the code book are publicly and freely available without restriction at https://dash.nichd.nih.gov/study/226675. The study investigators and the Data Coordinating and Analysis Center have copies of the entire database but cannot release that version to outside investigators owing to the permissions granted by the participants during the consent process and to protect participant confidentiality.

Funding

This study was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) (R01 HD102313 to LMB and AIN); RTI International (U10 HD063036); Case Western Reserve University (U10 HD063072); Columbia University (U10 HD063047); Indiana University (U10 HD063037); University of Pittsburgh (U10 HD063041); Northwestern University (U10 HD063020); University of California Irvine (U10 HD063046); University of Pennsylvania (U10 HD063048); and University of Utah (U10 HD063053). Support was also provided by respective Clinical and Translational Science Institutes to Indiana University (UL1TR001108) and University of California Irvine (UL1TR000153). The study sponsor had no role in the study design; collection, analysis, and interpretation of data; writing the report; or the decision to submit the report for publication.

Author disclosures

The authors report no conflicts of interest.

Appendix A. Supplementary data

The following is the Supplementary data to this article: