Discrimination and Health Among First-Generation Hispanic/Latinx Immigrants: the Roles of Sleep and Fatigue

Abstract

Introduction

A growing literature documents the associations between discrimination and health. Emerging evidence suggests that among Hispanic/Latinx immigrants, discrimination leads to the deterioration of health outcomes over time. While sleep has been proposed as an important mediator of the relationship between discrimination and health, few studies have explicitly investigated this pathway, particularly among Hispanic/Latinx populations.

Objective

To investigate the relationships between racial/ethnic discrimination, sleep, and physical and mental health among Hispanic/Latinx immigrants in the USA.

Data and Methods

Using data from a parent study of first-generation Hispanic/Latinx immigrants in the southeastern USA, we conducted sequential mediation analyses using the bootstrapping method to investigate whether self-reported sleep duration, sleep quality, and fatigue mediate the relationship(s) between self-reported discrimination, as measured by the discrimination subscale of the Riverside Acculturative Stress Inventory, and self-reported physical and mental health.

Results

Nocturnal awakenings, fatigue, and sleep quality were statistically significant sequential mediators of the relationship between discrimination and physical health (b = −.001, SE = .001, CI [−.0027, −.0001]); fatigue alone also mediated this relationship (b = −.01, SE = .01, CI [−.0279, −.0003]). Nocturnal awakenings, fatigue, and sleep quality were also significant sequential mediators of the relationship between discrimination and mental health (b = −.001, SE = .001, CI [−.0031, −.0001]).

Conclusion

Sleep and fatigue play an important role in linking discrimination and health among first-generation Hispanic/Latinx immigrants. The development and implementation of interventions that focus on reducing fatigue among this population could mitigate the effects of unfair treatment on health outcomes.

Introduction

A growing literature documents the associations between discrimination and adverse mental and physical health outcomes [1, 2]. Discrimination, or unfair treatment based on group membership or identity (e.g., race) [3], which can take a variety of forms, affects health through multiple channels. For example, institutional discrimination (e.g., policies and practices) likely impacts health indirectly via unequal access to housing, education, and other economic resources, while interpersonal discrimination likely has a more direct impact on health in the form of physiological and psychological stress reactions [3, 4]. While Hispanic/Latinx immigrants often have better average health outcomes than their non-immigrant counterparts, emerging evidence points to both institutional and interpersonal discrimination as important factors leading to the deterioration of health outcomes over time within this population subgroup [2, 5, 6]. Understanding how racial discrimination affects the health of Hispanic/Latinx immigrants is particularly important because this group has experienced a stark increase in racialized xenophobia during the 45th presidential administration (due to the former administration’s use of vehement anti-immigrant rhetoric) [2, 7, 8]. However, researchers know little about the specific mechanisms by which discrimination adversely impacts health outcomes among this population. While initial evidence from other racial/ethnic groups has suggested that stress plays an important role, this research remains nascent. Addressing these knowledge gaps is an important first step in reducing the discrimination-related health disparities experienced by the largest immigrant subgroup in the United States [9, 10].

One understudied mechanism that may link discrimination and health is sleep. Both insufficient sleep and poor-quality sleep negatively affect physical and mental health, leading to increased risks of cardiovascular disease, type 2 diabetes, arthritis, obesity, stroke, depression, and premature mortality [4, 11–15]. Insufficient sleep also causes fatigue (i.e., low energy or excessive tiredness), which in turn can lead to a decline in physical activity, diminished concentration, and other adverse health outcomes [16–20]. The most recent available data from the National Health Interview Survey (2014–2018) show that approximately 32% of survey respondents report getting an insufficient amount of sleep (i.e., less than 7 hours of sleep per 24-hour period) and approximately 15.6% report feeling exhausted (i.e., fatigue) either daily or most days in the 3 months prior to the survey [21]. These sleep-related outcomes vary significantly by race/ethnicity: American Indian/Alaska Native (36.0% and 21.9%) and non-Hispanic (NH) Black individuals (41.4 and 15.9%) experience among the highest rates of insufficient sleep and fatigue, while NH White (30.5 and 16.1%), NH Asian (32.3 and 9.3%), and Hispanic individuals (32.7 and 13.7%) experience among the lowest rates [21].

In general, the extant evidence indicates that Hispanic/Latinx immigrants experience sleep-related outcomes that are, on average, as good as if not better than those of their US-born counterparts. These patterns are consistent with a body of research findings showing that foreign-born Hispanic/Latinx populations generally have better health-related outcomes compared to their US-born counterparts, despite being more socioeconomically disadvantaged (i.e., the Hispanic paradox) [22]. A recent review [23] found that some studies showed that Mexican immigrants were significantly less likely than US-born Mexicans to experience habitual short sleep (e.g., <7 hours) [24], while others found no significant nativity differences in the likelihood of short sleep among Hispanic/Latinx participants after accounting for demographic and other health-related characteristics [25, 26]. Similarly, Hispanic/Latinx immigrants report levels of sleep quality (i.e., insomnia-based sleep complaints) that are similar to [25] or better than [24] their US-born Hispanic/Latinx counterparts, although these advantages may be more concentrated among male immigrants [24]. Finally, there may be language-based differences in sleep outcomes: studies have found that compared to individuals who speak only Spanish at home, those who speak increasing levels of English experienced poorer sleep-related outcomes, even after controlling for immigrant status [24, 27].

The existing literature often attributes the erosion of immigrant health advantages, including sleep health, to acculturation (i.e., the adoption of negative host country behaviors) [25]. However, like other forms of stress, discrimination may contribute to the erosion of sleep health advantages among Hispanic/Latinx immigrants through both physiological and psychological pathways. First, experiences of discrimination can lead to the dysregulation of both the sympathetic-adrenal-medullary axis (SAM axis) and the hypothalamic-pituitary-adrenal axis (HPA axis). Prolonged activation in the SAM axis can elevate levels of norepinephrine in the brain, which in turn can enhance the amygdala functions (i.e., processing emotions, particularly negative ones) and impair the prefrontal cortex functions (i.e., reasoning and critical thinking) and therefore heighten attentional biases toward negative information [28]. Thus, one psychological consequence of discrimination is increased worry and rumination, which can impact a person’s ability to fall asleep and stay asleep [29]. Increased HPA activation can also inhibit sleep, lead to sleep disruptions, and increase cortisol levels [30]. While cortisol may not be the proximate cause of sleep disturbances, elevated cortisol levels are thought to be an important indicator of increased central nervous system activity and nocturnal awakenings [30].

These proposed pathways align with the findings reported in Slopen et al.’s systematic review showing that higher levels of self-reported discrimination are associated with shorter duration of sleep, poorer sleep quality (ability to fall asleep and nocturnal awakenings), less efficient sleep (ratio of time spent asleep to time spent in bed), and greater fatigue across racial and ethnic groups, including Hispanic/Latinx populations [4, 11]. More recent studies have found that racial/ethnic discrimination is linked to poorer sleep—including short sleep duration, high sleep-onset latency, and poor sleep quality—and fatigue in both adolescents and adults [31–35]. Studies examining racial/ethnic discrimination and sleep outcomes among exclusively Hispanic samples have also found that these experiences of discrimination are associated with poorer sleep outcomes—including short and long sleep durations and high levels of daytime sleepiness—and fatigue [36–42]. (For an important recent exception, see Alcántara and colleagues (2019).)

Despite these informative results, few studies have explicitly examined sleep as a mediator of the relationship between discrimination and health outcomes. In a review of the existing literature, we identified four studies that investigated the relationships between discrimination, sleep, and health in multiracial samples (three of the samples included five racial/ethnic groups and the fourth included two; the following groups were included in at least one of the samples: NH White, NH Black, Hispanic/Latinx, NH Asian and Pacific Islander, and Native American participants); all four found that sleep quality mediated all or part of the relationship(s) between discrimination and mental and/or physical health [43–46]. However, with two exceptions [43, 46], these studies examined sleep as a mediator of the relationship between general discrimination, rather than racial/ethnic discrimination, and health. None examined the relationships between discrimination, sleep, and health among Hispanic/Latinx populations in particular.

To our knowledge, only two prior studies have investigated whether sleep mediates the relationship between racial/ethnic discrimination and health outcomes in Hispanic/Latinx populations; furthermore, these studies examined mental health outcomes (specifically depressive symptoms) but not physical health outcomes [37, 42]. In the first study, Pichardo et al. [37] found that sleep quality, but not sleep efficiency, mediated the relationship between experiences of ethnic discrimination and depressive symptomology in Latinx college students. However, because the study sample was small and included only college students, the findings have certain limitations. For example, the authors could not determine whether the relationships between discrimination, sleep, and health varied by nativity. Furthermore, because Latinx college students are not representative of the broader Latinx population, the findings might not be generalizable. That is, college students are, on average, physically healthier than the general population [47, 48]. In addition, relative to first-generation Latinx college students, the wider population of first-generation Latinx immigrants might not be as familiar with the nature of the structural, institutional, and interpersonal forms of discrimination in the United States. Thus, they may not interpret their experiences of unfair treatment as discrimination [49].

In the second study, based on data from 168 adult Hispanic immigrants in Utah, Steffen and Bowden (2006) found that sleep disturbances mediated the relationship between high levels of perceived racism and symptoms of depression. While these findings are informative, we argue that because of changes in the composition of the Hispanic/Latinx immigrant population over the past 15 years [50], as well as variation in social and geographic contexts, it is important to examine the relationships between discrimination, sleep, and health in more contemporary samples from additional regions of the United States.

In the current study, we build on prior research to investigate the relationships between racial/ethnic discrimination, sleep, and physical and mental health among Hispanic/Latinx immigrants in the USA. We hypothesize that sleep and fatigue mediate the association(s) between racial/ethnic discrimination and poor physical and mental health outcomes. Specifically, we hypothesize that experiencing racial/ethnic discrimination leads to fragmented sleep, and subsequently to poor sleep quality. In turn, poor sleep quality increases fatigue, which leads to subsequent poor health outcomes.

Methods, Data, and Analysis Plan

The present study analyzes data from a larger pilot study that examined the mechanisms linking racial/ethnic discrimination and health outcomes. Over a 12-month period (2018–2019), the parent study recruited 231 self-identified Hispanic/Latinx individuals age 18 and older who were born outside the continental USA and were currently living in Central Virginia. All participants provided informed consent. We worked closely with a community consultant with a strong record of recruitment success in the target community. The consultant helped the research team secure multiple data collection opportunities within the Hispanic/Latinx community and advertised the study via social media and word of mouth. Participants were recruited in a variety of venues, including churches (e.g., at a Latin American heritage event), libraries (e.g., at mobile Mexican consulate events), and neighborhood gathering spots (e.g., mobile home communities and hair salons). The study team also posted flyers around the university campus and in community settings such as grocery stores catering to Hispanic/Latinx customers. The study protocol received approval from the Institutional Review Board at Virginia Commonwealth University.

After bilingual research assistants administered the informed consent protocol, participants were invited to complete a survey either on a laptop or using a paper version, in either English or Spanish. (Note: The survey questions were developed with feedback from our bilingual study team members and our community consultant. We also used the services of professional translators to ensure that the questions were valid in both English and Spanish and had our bilingual team members review the questions after translation.) The survey captured demographic characteristics, self-reported health outcomes, and experiences of racial/ethnic discrimination. For the current analyses, we excluded 17 participants who had missing information on health- or sleep-related outcomes or relevant demographic characteristics. Table 1 summarizes participant characteristics.

Table 1.

Participant characteristics

	M (SD) or Freq. (%) (included in the analyses n = 214)	M (SD) or Freq. (%) (excluded from the analyses n = 17)
Age	38.88 (11.94)	38.65 (11.40)
Gender	Missing n = 1
Women	145 (67.8%)	9 (52.9%)
Men	68 (31.8%)	8 (47.1%)
Age of immigration	24.08 (11.74)	24.69 (10.26)
Years in the USA	15.39 (9.80)	14.75 (3.97)
Income	Missing n = 57	Missing n = 1
$10,000 or less	24 (11.2%)	2 (11.8%)
$10,001-$20,000	26 (12.1%)	7 (41.1%)
$20,001-$30,000	28 (17.7%)	4 (23.5%)
$30,001-$40,000	29 (13.4%)	2 (11.8%)
$40,001-$50,000	18 (8.4%)	1 (5.9%)
$50,001 or above	22 (10.3%)	1 (5.9%)
Marital status	Missing n = 3
Married	102 (47.7%)	4 (23.5%)
Not married	109 (50.9%)	13 (76.5%)
US citizenship status	Missing n = 6
US citizen	52 (24.3%)	2 (11.8%)
Permanent resident	18 (8.4%)	--
Temporary immigrant	138 (64.5%)	15 (88.2%)
Country/region of origin	Missing n = 2
Mexico	123 (57.5%)	16 (94.1%)
Puerto Rico	20 (9.3%)	--
Cuba	2 (0.9%)	--
Central or South America	56 (26.2%)	1 (5.9%)
Other	11 (5.1%)	--

Measures

In the following section, we review the measures included in the current analysis of the survey data.

Self-Reported Health

The main health-related outcomes of interest were self-rated physical and mental health. Respondents were asked: “In general, how would you rate your physical health?” A similar question asked respondents to rate their mental health. Respondents could choose from five options (poor, fair, good, very good, and excellent), which we transformed into a 5-point scale (range=1–5), with higher numbers indicating better health. Researchers have consistently found that self-rated physical health is linked to both subjective and objective measures of morbidity, as well as mortality [51–54]. Similarly, self-rated mental health is associated with both mental morbidity, including depressive symptoms and non-specific psychological distress, and activity limitations [55, 56].

Sleep

Duration

To capture sleep duration, the survey asked the following open-ended question: “During a regular week (including both days you work and days you do not work), how many hours do you sleep on average?” Participants wrote in the average number of hours they slept per night.

Difficulty Falling Asleep

To assess problems falling asleep the survey asked: “During a regular week (including both days you work and days you do not work), how often did you have difficulties falling asleep on average?” Response options included never, rarely, sometimes, often, and frequently, which were transformed into a 5-point scale (range=1–5), with higher numbers indicating more difficulty.

Waking up at Night

To assess how often participants wake up during the night (sometimes called nocturnal awakenings), the survey asked: “During a regular week (including both days you work and days you do not work), about how many times do you wake up during the night on average?” Respondents wrote in the number of awakenings, which we coded as a continuous variable.

Difficulty Going Back to Sleep

The survey also asked: “During a regular week (including both days you work and days you do not work), how difficult is it to go back to sleep again after you wake up on average?” Response options included very easy, somewhat easy, somewhat difficult, and very difficult, which were transformed into a 4-point scale (range=1–4), with higher numbers indicating greater difficulty.

Quality

To measure the self-reported quality of participants’ sleep, the survey asked: “In general, how would you describe your sleep quality?” Response options included poor, fair, good, very good, and excellent; responses were transformed into a 5-point scale (range=1–5), with higher numbers indicating better self-reported sleep quality.

Fatigue

To capture participants’ experiences of fatigue, the survey asked: “In general, how tired do you feel during the day?” Participants could choose one of four response options: not at all tired, somewhat tired, very tired, and extremely tired, which were transformed into a 4-point scale (range=1–4), with higher numbers indicating higher levels of fatigue.

Racial/Ethnic Discrimination

We used the discrimination subscale of the Riverside Acculturation Stress Inventory (RASI; Benet-Martinez (2003)) to capture racial/ethnic discrimination [57, 58]. The RASI has demonstrated strong psychometric properties (e.g., reliability, validity) in prior work [58]. The discrimination subscale contains three items: “I have been treated rudely or unfairly because of my Latino/Latina background,” “I feel that people very often interpret my behavior based on their stereotypes of what Latinos/Latinas are like,” and “I have felt discriminated against by Americans because of my Latino/Latina background.” Each item is rated on a 5-point scale ranging from 1=“Strongly disagree” to 5=“Strongly agree.” We summed the scores for the three items; higher numbers indicate higher levels of perceived discrimination. The internal reliability of the discrimination subscale was α = .88.

Other Covariates

Finally, we included a set of demographic characteristics that previous studies have found are linked to mental and physical health among immigrants, including age, sex, household income, years of residence in the USA, and US citizenship status [59–61].

Analysis Plan

We first calculated descriptive statistics, including means, frequencies, and standard deviations for the variables in the current study. We then conducted two sets of preliminary analyses before completing the main hypothesis testing. In the first set, we calculated product-moment correlations to identity potential mediators of the link(s) between discrimination and health.

In the second set of analyses, we calculated product-moment correlations between demographic characteristics and physical/mental health as well as a one-way ANOVA with citizenship status as a between-subject variable; the goal of these analyses was to identify covariates that were associated with physical health, mental health, or both.

The main hypothesis testing was done via mediation analyses based on the bootstrapping method with N = 5000 resamples in PROCESS [62, 63] for each health outcome separately. Specifically, we tested our hypothesis by conducting a sequential mediator model (i.e., discrimination → waking up at night → sleep quality → fatigue → physical/mental health) using Model 6 in PROCESS. We also conducted two sets of follow-up analyses to explore additional questions. First, we conducted parallel mediator models using PROCESS Model 4 to explore whether the association between racial/ethnic discrimination and each health outcome was mediated by fatigue, sleep quality, and waking up at night independently. Second, we also explored all possible patterns of the sequences among the three mediators using Model 6 in PROCESS (see Table 2). We also controlled for relevant covariates of interest (gender, income, and US citizenship status). In all mediation analyses, discrimination, fatigue, sleep quality, and waking up at night were grand-mean-centered, and sex (reference group=female) and US citizenship status (reference group=citizen, permanent resident, or immigrant) were dummy-coded.

Table 2.

Sequences of the three sleep-related mediators

Predictor				Mediators				Outcome
Discrimination	→	1. Fatigue	→	Sleep quality	→	Waking at night	→	Physical/mental health
		2. Fatigue	→	Waking at night	→	Sleep quality
		3. Sleep quality	→	Fatigue	→	Waking at night
		4. Sleep quality	→	Waking at night	→	Fatigue
		5. Waking at night	→	Fatigue	→	Sleep quality
		6. Waking at night	→	Sleep quality	→	Fatigue

Results

Descriptive Statistics

Table 3 presents the means, standard deviations, and partial correlations for the primary variables. Physical and mental health were significantly associated with one another (r = .52, p < .001), such that participants who reported better physical health also reported better mental health. Waking up at night was significantly and negatively associated with physical health (r = −.19, p = .026) but not mental health (r = −.08, p = .338). Table 1 presents the means and standard deviations for participant characteristics.

Table 3.

Means, standard deviations, and partial correlations among the primary variables

		1	2	3	4	5	6
1. Discrimination		--
2. Physical health		−.18*	--
3. Mental health		−.13	.52***	--
4. Fatigue		.25**	−.33***	−.26**	--
5. Sleep quality		−.16†	40***	.43***	−.22**	--
6. Waking at night		.15†	−.19*	−.08	.18*	−.17*	--
	M	9.45	2.66	3.38	2.19	2.90	1.85
	(SD)	(3.94)	(.93)	(1.03)	(.93)	(1.04)	(1.67)

^†p < .10,

^*p < .05,

^***p < .001.

Gender, income, marital status, and US citizenship status were controlled

Preliminary Analyses

The preliminary analyses (calculating product-moment correlations to identify discrimination/health mediators) revealed that fatigue was significantly associated with racial/ethnic discrimination (r = .23, p = .001), physical health (r = −.35, < .0001), and mental health (r = −.27, p < .0001). Of the five sleep-related measures, only self-reported sleep quality was significantly associated with racial/ethnic discrimination (r = −.14, p = .045), physical health (r = .42, p < .001), and mental health (r = .43, p < .001). The frequency of “night awakenings” was significantly associated with physical health (r = −.17, p = .013) but not mental health (r = −.09, n.s.). Thus, only fatigue, sleep quality, and waking up at night were included in the main hypothesis testing.

The results of the second set of preliminary analyses (product-moment correlations and/or one-way ANOVA tests to identify the key demographic characteristics associated with physical and/or mental health) indicated that both gender and income were significantly and positively associated with physical health only (r = .17, p = .013 and r = .18, p = .024, respectively). The ANOVA results revealed that US citizenship status significantly predicted physical health but not mental health, F(2,201) = 4.79, SE = 4.03, p = .009. Specifically, participants with permanent residency reported significantly better physical health (M = 3.28, SD = .90) than both participants with US citizenship (M = 2.71, SD = 1.01) and participants with temporary immigrant status (M = 2.57, SD = .89); the latter two groups were not statistically different from one another. Finally, neither age nor years of residence in the USA was associated with physical or mental health. Thus, only gender, income, and US citizenship status were included in the main hypothesis testing models as covariates.

Main Hypothesis Testing

Physical Health

While the total effect of racial/ethnic discrimination on physical health (Fig. 1A) was statistically significant (c: b = −.06, SE = .02, p < .01, CI [−.0960, −.0153]), the direct effect was only marginally significant (c’: b = −.03, SE = .02, p = .0890, CI [−.0690, .0050]). Prior research suggests that the presence of a direct effect between independent and dependent variables is not necessary to establish a significant mediation model [64, 65]; thus, we proceeded to investigate whether racial/ethnic discrimination had an indirect effect on physical health via the sleep mediators in the predicted sequence (i.e., discrimination → waking up at night → sleep quality → fatigue → physical health). Inconsistent with our prediction(s), the link between discrimination and physical health was not mediated by waking up at night, sleep quality, and fatigue sequentially (a¹*d²¹*d³²*b³: b = −.0004, SE = .0004, CI [−.0004, −.0001].

Fig. 1.

A, B Mediation analyses of discrimination predicting participants’ self-reported physical and mental health with gender, income, marital status, and US citizenship status as covariates. Note. † p < .10, * p < .05, *** p < .001 Note. † p < .10, * p < .05, *** p < .001.

Mental Health

Neither the total effect of discrimination on mental health (c: b = −.03, SE = .02, p = .14, CI [−.0794, .0110]) nor the direct effect of discrimination on mental health (c’: b = −.02, SE = .02, p = .48, CI [−.0576, .0275]) was statistically significant. The sequential mediation among discrimination, waking up at night, sleep quality, fatigue, and mental health was not significant, either (a¹*d²¹*d³²*b³: b = −.0004, SE = .0004, CI [−.0014, −.0001].

Follow-up Analyses

Physical Health

Given the lack of support for our hypothesis, we conducted two sets of follow-up analyses. First, we explored whether the three sleep mediators independently (as opposed to sequentially) explain the association between discrimination and physical health. A parallel mediation analysis (Model 4) revealed that only fatigue was a significant mediator (a¹*b¹: b = −.01, SE = .008, CI [−.0300, −.0015]). Second, we explored five additional possible indirect pathways based on the unique combinations of the three mediators (Table 4). Only one indirect effect (discrimination → waking up at night → fatigue → sleep quality → physical health, Table 4, Pathway 6) was statistically significant (a¹*d²¹*d³²*b³: b = −.001, SE = .001, CI [−.0027, −.0001]). These results suggest that higher levels of self-reported discrimination are positively associated with the number of nocturnal awakenings, which, in turn, is associated with an increased level of fatigue. Greater fatigue is then associated with poorer sleep quality, which is related to poorer physical health. In addition, the association between discrimination and physical health was mediated by fatigue alone (a²*b²: b = −.01, SE = .01, CI [−.0279, −.0003]) and was sequentially mediated by fatigue and then sleep quality (a²*d³²*b³: b = −.01, SE = .003, CI [−.0139, −.0006]). These results suggest that fatigue might be a particularly critical mediator of the association between discrimination and physical health. Figure 1A summarizes the results of the mediation analysis for all six potential pathways.

Table 4.

Summary of the indirect effects of discrimination on physical and mental health through different sequential combinations of fatigue, sleep quality, and waking at night

	Mediation sequence	b	SE	LLCI	ULCI
Physical health
1.	Waking at night → sleep quality → fatigue	−.0004	.0004	−.0014	.0001
2.	Fatigue → sleep quality → waking at night	−.0001	.0003	−.0008	.0003
3.	Fatigue → waking at night → sleep quality	−.0003	.0005	−.0015	.0004
4.	Sleep quality → fatigue → waking at night	−.0001	.0002	−.0008	.0002
5.	Sleep quality → waking at night → fatigue	−.0001	.0002	−.0006	.0001
6.	Waking at night → fatigue → sleep quality	−.0010	.0007	−.0027	−.0001
Mental health
1.	Waking at night → sleep quality → fatigue	−.0004	.0004	−.0014	.0001
2.	Fatigue → sleep quality → waking at night	.0000	.0003	−.0006	.0008
3.	Fatigue → waking at night → sleep quality	−.0004	.0006	−.0020	.0005
4.	Sleep quality → fatigue → waking at night	.0000	.0002	−.0004	.0005
5.	Sleep quality → waking at night → fatigue	−.0001	.0002	−.0005	.0001
6.	Waking at night → fatigue → sleep quality	−.0012	.0008	−.0031	−.0001

Order #1 is the hypothesized order based on prior research and theories. Orders #2–6 are tested as part of exploratory analyses. Gender, income, marital status, and US citizenship status were included in all mediation analyses

Mental Health

Given the lack of evidence supporting the total effect of discrimination on mental health, none of the mediators came out as a significant mediator in a parallel mediation analysis. Somewhat surprisingly, however, one of the five remaining sequential mediations became significant: discrimination → waking up at night → fatigue → sleep quality → mental health (Table 4, Pathway 6) (a¹*d²¹*d³²*b³: b = −.001, SE = .001, CI [−.0031, −.0001]). These results indicate that higher levels of reported discrimination predict more frequent waking during the night, which further predicts a higher level of fatigue. Greater fatigue, in turn, predicts poorer sleep quality, which is associated with poorer mental health. Consistent with the results for physical health, the association between discrimination and mental health was also sequentially mediated by fatigue and then sleep quality (a²*d³²*b³: b = −.01, SE = .004, CI [−.0168, −.0009]). However, unlike the results for physical health, the association between discrimination and mental health was not mediated by fatigue alone (a²*b²: b = −.01, SE = .01, CI [−.0252, .0014]). Figure 1B summarizes the results of the mediation analysis predicting mental health.

Discussion

Prior research has documented associations between discrimination and physical and mental health [3]. However, few studies have explicitly examined how sleep and fatigue contribute to these relationships among Hispanic/Latinx adults. In the present study, we investigated the roles of sleep duration, sleep quality, and fatigue as potential mechanisms linking experiences of discrimination and physical and mental health in a sample of first-generation Hispanic/Latinx immigrants. The findings are threefold: First, the results showed that nocturnal awakenings, fatigue, and self-reported sleep quality constitute a critical pathway linking discrimination and self-reported physical health, accounting for approximately 13% of the total effect of perceived discrimination on self-reported physical health. That is, respondents who reported higher levels of racial/ethnic discrimination were also more likely to report more nocturnal awakenings, which are associated with greater fatigue and poorer sleep quality. In turn, fatigue and poor self-reported sleep quality were linked to poor physical health. In addition, fatigue alone mediated the relationship between discrimination and physical health, indicating that fatigue plays a particularly important role in the discrimination-health link and might be an important intervention target point. Second, we did not detect a direct relationship between discrimination and self-reported mental health. However, as with physical health, the findings indicated that nocturnal awakenings, fatigue, and self-reported sleep quality mediated the relationship between discrimination and mental health. Third, the analyses produced no evidence that sleep duration mediated the relationships between discrimination and self-reported health. To our knowledge, this study is one of the first to investigate these relationships among first-generation immigrants.

The finding that nocturnal awakenings, fatigue, and self-reported sleep quality are important mediators of the relationship between discrimination and self-rated physical health is consistent with prior research findings from both cross-sectional and longitudinal studies showing that discrimination disrupts sleep continuity (i.e., increases nocturnal awakenings), and that sleep continuity affects fatigue and sleep quality [4]. Furthermore, in a recent longitudinal study using data from the Midlife in the United States Study, Hisler and Brenner (2019) found that fatigue explains the sleep-related pathway linking discrimination and physical health.

We also found that fatigue alone mediated the relationship between perceived discrimination and physical health. The racial battle fatigue framework proposes that constantly coping with racial microaggressions places mental, emotional, and physical strain on the body, thus resulting in daily fatigue [66, 67]. Findings from the Hispanic Community Health Study/Study of Latinos, for example, indicate that reporting a high level of ethnic discrimination is linked to greater fatigue among Latinx adults [38]. Similarly, several studies have linked experiencing a high level of discrimination to fatigue in samples of Black, White, and/or Latinx adults [14, 35, 68].

The current finding that discrimination has adverse effects on self-reported mental health via two pathways—fatigue and sleep quality as well as waking up at night, fatigue, and sleep quality—is partially consistent with prior research using both multiethnic and Hispanic/Latinx-only samples. Some prior studies have found direct relationships between discrimination and mental health [43, 44], while other studies, including the present one, have found that the direct relationship was not statistically significant but sleep mediated the discrimination-health link [37]. For example, in a study of Latinx college students, Pichardo (2020) found that racial/ethnic discrimination was indirectly associated with depressive symptoms through racism-related vigilance and sleep quality. While we were unable to measure racism-related vigilance in the present study, it may lead to higher levels of stress and anxiety, and in turn cause sleep disturbances, which are strongly associated with depression and other mental health outcomes. Similar to Steffen and Bowden (2006), we found that sleep disturbances mediated the relationship between discrimination and health (although the coefficient for night awakenings was of marginal significance). Steffen and Bowden suggested that experiencing higher levels of discrimination may lead to intrusive thoughts and, eventually, sleep disruption.

Finally, the finding that sleep duration was not significantly related to either discrimination or self-reported health is consistent with some prior studies that have found that sleep quality is more strongly associated with health outcomes than sleep duration [69, 70]. The findings also align with those of prior studies showing that sleep quality, but not duration, mediates the relationship between discrimination and health [4, 42] and that overt discrimination is significantly associated with sleep quality but not sleep duration [71]. In a partial exception, Yang and Park (2015) found that together sleep quality and duration accounted for 15–25% of the adverse effect of perceived discrimination. However, they noted that sleep quality was a far more salient mediator of the relationship between discrimination and health. Although the limitations of the data used in the current analysis prevented us from examining the underlying reasons for the lack of a statistically significant effect of sleep duration, we propose that using the overall duration of sleep as a measure does not capture the ways in which discrimination may disrupt key components of sleep architecture. For example, studies using objective measures of sleep monitoring have found that higher levels of self-reported discrimination are linked to decreases in sleep continuity and less time spent in stages 3 and 4 of restorative slow-wave sleep (SWS) [11, 68, 72, 73]. These stages of sleep are particularly important for cortisol inhibition, which is responsible for metabolism, energy, and immune function [11]. Although the focal measure of self-reported sleep quality is also incomplete, individual perceptions of sleep may be relatively better assessments of these components of sleep architecture.

Limitations and Future Research Directions

It is important to acknowledge several limitations of the present study. First, the study sample was drawn from a convenience sample of community residents, and any resulting bias might have impacted the study’s findings. Notably, however, we were able to engage a fairly sizable sample of Hispanic/Latinx immigrants during a period of heightened immigration enforcement and xenophobia that could have posed significant challenges to study recruitment. Importantly, while half of the Hispanic/Latinx population resides in traditional destination states such as Texas and California [74], the Hispanic/Latinx population living in the US South has experienced rapid growth relative to the populations in other areas (e.g., the West Coast) and comes from a significantly more varied group of countries of origin. For example, Virginia has experienced a 31% increase in its Hispanic/Latinx population since 2010 [75] and approximately 46% of Virginia’s Hispanic/Latinx population is foreign-born—one of the highest percentages in the USA [76]. As such, the study population can provide new insights into how discrimination and sleep contribute to the health disparities experienced by Hispanic/Latinx immigrants.

Second, the study used measures of sleep and health that were subjective rather than objective. Prior research has found that the relationships between discrimination and sleep outcomes vary depending on whether researchers rely on self-reports or use objective measures (e.g., sleep actigraphy) [4]. Although the self-rated mental health item is linked to psychological distress, it is not a substitute measure for clinically elevated symptoms and does not capture information on different types of mental health symptoms (e.g., anxiety vs. depression vs. trauma) [55, 56]. Furthermore, while self-rated health has robust links to physical morbidity and mortality, scholars such as Viruell-Fuentes (2011) have argued that the most common Spanish translation of the “fair” option in the self-rated health scale (i.e., “regular” in Spanish) may inadvertently cause Spanish-speaking individuals to rate their health more poorly than their English-speaking counterparts [77]. However, this linguistic difference is likely more problematic in studies comparing outcomes between US-born and foreign-born respondents than in the present study, which compares subgroups of immigrants; furthermore, there was no evidence of variation in interpretations of the self-rated health scale by level of self-reported discrimination.

Third, we used cross-sectional data, which did not allow us to definitively assess temporal ordering and thus we could not make causal inferences. This may explain, for example, why the models identified a statistically significant path from fatigue to sleep quality rather than the more logical path from sleep quality to fatigue. Future longitudinal and/or prospective studies that integrate multiple components of sleep and use both objective and subjective measures could facilitate causal inferences and provide evidence to inform appropriate interventions [4].

Fourth, given data limitations we were unable to fully explore how other pertinent factors affect the identified relationships between discrimination, sleep, and health among Hispanic/Latinx immigrants. While prior research has identified important differences in mental health, physical health, and sleep patterns among Hispanic immigrants by country of origin [26, 78], sample size limitations prevented us from exploring these differences in the present study. Furthermore, given the need to streamline the survey to maximize response rates, we chose not to collect information on other relevant characteristics that could help explain the discrimination-sleep link [4]. For example, prior research has suggested that daytime dysfunction, or struggling to remain alert and motivated during the day, may be an important contributing factor [42, 44]. Specifically, individuals who experience fatigue may have difficulty staying alert and motivated during the daytime and thus have limited capacity to regulate health behaviors, which in turn, leads to worse health outcomes [42, 44]. In a related study, Steffen and Bowden (2006) identified daytime dysfunction as a key factor linking discrimination and somatic symptoms. Future studies should integrate this information to better characterize the links between discrimination, sleep, and health among immigrants.

Finally, the discrimination measure captured only interpersonal experiences of racial/ethnic discrimination, which could underestimate the total association(s) between discrimination, sleep, and health. Other types of discrimination (e.g., based on LGBTQ, disability, gender, and immigrant status, as well as other characteristics such as language use and national identity) might have additive and interactive effects on sleep and health outcomes over time [2, 4]. We chose to focus on racial/ethnic discrimination because the current political climate has resulted in Hispanic/Latinx immigrants experiencing increased racialized discrimination [2, 7]. Given that initial studies using more general forms of discrimination and longitudinal data [44] have found a relationship between discrimination, sleep, and health, future research should examine the effects of discrimination based on other statuses.

Conclusion

This study makes an important contribution to the nascent literature on discrimination, fatigue, sleep, and health. The findings demonstrate the particularly important role of fatigue in mediating the link between discrimination and health among Hispanic/Latinx immigrants. Furthermore, they suggest that rather than focusing on acculturation-related behavioral changes as the primary source of immigrant health deterioration, it is critical to better understand how unfair racialized treatment shapes one of the most crucial biological processes—sleep. In the long run, addressing structural racism and xenophobia will play an important role in addressing health declines among Hispanic/Latinx immigrants. However, in the short run, interventions that address fatigue among first-generation Hispanic/Latinx immigrants who experience discrimination could also mitigate the effects of unfair treatment and help reduce associated health disparities.