Employment status and the association of sociocultural stress with sleep in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL)

ABSTRACT

Study Objectives

We examined the association of sociocultural stress severity (i.e. acculturation stress, ethnic discrimination) and chronic stress burden with multiple dimensions of sleep in a population-based sample of US Hispanics/Latinos. We also explored whether employment status modified stress-sleep associations.

Methods

We conducted survey linear regressions to test the cross-sectional association of sociocultural stress severity and stress burden with sleep dimensions using data collected between 2010 and 2013 from individuals who participated in both the Hispanic Community Health Study/Study of Latinos Sueño and Sociocultural Ancillary studies (N = 1192).

Results

Greater acculturation stress (B = 0.75, standard error [SE] = 0.26, p < .01) and chronic psychosocial stress burden (B = 1.04, SE = 0.18, p < .001) were associated with greater insomnia symptoms but were not associated with actigraphic measures of sleep. Ethnic discrimination was not associated with any of the sleep dimensions. The association of acculturation stress with insomnia severity was greater in unemployed (B = 2.06, SE = 0.34) compared to employed (B = 1.01, SE = 0.31) participants (p-interaction = .08).

Conclusions

Acculturation stress severity and chronic stress burden are important and consistent correlates of insomnia, but not actigraphically measured sleep dimensions. If replicated, future research should test whether interventions targeting the resolution of sociocultural stress improve sleep quality in Hispanics/Latinos.

Statement of Significance.

While sleep is increasingly recognized as an important public health issue, Hispanic/Latino sleep health remains understudied. This study is among the first to examine the association of sociocultural stressors and chronic stress burden with subjectively and objectively measured sleep dimensions in Hispanics/Latinos, and to explore if employment status modifies these associations. Greater acculturation stress and chronic psychosocial stress burden were associated with greater insomnia symptoms, but not with actigraphic measures of sleep. Ethnic discrimination was not associated with insomnia or actigraphic measures of sleep. The effect of acculturation stress on insomnia symptoms may be stronger in unemployed compared to employed participants. If replicated, future research should test whether behavioral interventions targeting the resolution of acculturation stress improve sleep quality among Hispanics/Latinos.

Introduction

Sleep is increasingly recognized as an important behavioral and public health issue for all in the United States [1, 2]. Yet, Hispanic/Latino sleep health is understudied despite the fact that Hispanics/Latinos comprise 16.3% of the US population, and that sleep problems are prevalent among Hispanics/Latinos [3, 4]. In addition, while general psychosocial stress is a known determinant of self-reported poor sleep [5], it is unclear how exposure to specific sociocultural stress, that is, the stress linked to an individual’s social position or cultural identity (e.g. acculturation stress, discrimination stress), affects both subjectively and objectively measured sleep among Hispanics/Latinos. Further, whether and how upstream factors, such as an individual’s employment status, modify the stress–poor sleep disease course is unknown despite longstanding evidence of the effects of employment on health [6]. Greater understanding of these relationships could inform the development of targeted and contextually informed public health interventions to reduce stress and improve sleep in Hispanic/Latino communities.

Recent conceptualizations define sleep health as multidimensional and consisting of five dimensions, namely sleep satisfaction/quality, alertness, duration, timing, and efficiency, each of which can be measured across levels of analysis (subjective [self-report], behavioral, and physiological) [7]. To that end, there is both experimental and field research showing that nonspecific psychosocial stress is adversely associated with multiple dimensions of sleep, namely insomnia (an indicator of sleep satisfaction or quality), duration, and efficiency, when measured using either subjective (self-report) [8–12] or objective (behavioral, physiological) methods [8, 13–16]. However, studies testing the association of stress with objectively measured sleep have produced inconsistent findings. In some instances, an adverse association between stress and objectively measured sleep was observed [8, 13–16], and in other instances no statistically significant associations were found [16, 17]. Far less research has examined the stress–sleep association in racial/ethnic minority samples [12]. Together, these findings highlight the gaps in knowledge in the overall association of stress with objectively measured sleep, particularly among racial/ethnic minorities, and underscore the importance of sleep measurement in understanding the stress–sleep association.

With regard to sociocultural stress and sleep, a small but growing empirical literature indicates that exposure to various sociocultural stressors such as ethnic discrimination and acculturation-related stressors are linked to poor sleep [18, 19]. For example, a systematic review of 17 studies found that exposure to ethnic discrimination was associated with various indicators of poor sleep, with the most consistent results observed for indicators of sleep satisfaction measured using self-report [20]. Other research has found that proxy measures (e.g. US-born vs. foreign-born nativity status, greater English language proficiency) of acculturation [21], that is the individual process of adaptation and integration resulting from contact with an unfamiliar culture [22, 23], were associated with worse sleep among racial/ethnic communities [19, 24–27]. Our own work has shown that greater acculturation stress, defined as the psychological distress/worry associated with the multidimensional process of acculturation [21], and ethnic discrimination were independently associated with self-reported indicators of poor sleep, including insomnia, sleepiness, and short and long sleep duration in a large probability-based sample of Hispanic/Latino adults [28]. We did not, however, examine the association of sociocultural stress with objectively measured sleep parameters.

Importantly, the association of stress with sleep may be modified by employment status. According to prevailing models of insomnia pathogenesis, insomnia arises from the presence of predisposing, precipitating, and perpetuating factors [29]. Stress may serve as a precipitating factor for insomnia that is magnified in the context of unemployment. Specifically, unemployment may facilitate the development of maladaptive behaviors (e.g. excessive time in bed, daytime napping, rumination) and environmental responses, which in turn serve to perpetuate insomnia.

Most studies on stress and sleep have historically examined a single sleep dimension using only one assessment method (self-report, behavioral, or physiological). Yet, the correlation between objectively and subjectively measured sleep is weak to modest at best [30], including in Hispanic/Latino samples [31]. Further, most studies on the association of stress and objectively measured sleep dimensions have been conducted in experimental settings, leaving significant gaps in our understanding of the ecological association of stress with objectively measured sleep outside of the laboratory. Moreover, as discussed previously, there is notable discordance between existing findings on the association of stress with subjectively measured sleep versus objectively measured sleep. Additionally, most of this research has been conducted without a comprehensive assessment of sociocultural stress, sufficient attention to modifying factors such as employment status, or integration of racial/ethnic minority samples such as Hispanic/Latinos. Indeed, none of the extant published literature on sleep and psychosocial stress generally, and sleep and ethnic discrimination specifically, have examined these issues in a large probability-based sample of Hispanics/Latinos.

To address these limitations, we leveraged the comprehensive sleep and sociocultural assessment from the Sociocultural Ancillary Study (SCAS) and the Sueño ancillary study of the Hispanic Community Health Study/Study of Latinos (HCHS/SOL) to examine the cross-sectional associations of sociocultural stress severity and stress burden (acculturative stress, ethnic discrimination, and chronic stress) with insomnia and objectively measured sleep dimensions (i.e. sleep duration, sleep variability, sleep efficiency). We also explored if employment status modified the stress–sleep association. We focused on insomnia symptoms because of the increasing prevalence of insomnia among Hispanics/Latinos in the United States [32]. Temporal analyses using the National Health Interview Survey indicate that the age-adjusted prevalence of insomnia among Hispanic/Latino adults was 16.6% in 2002 compared to 19.3% in 2012 [32]. We hypothesized that higher scores on all three stress measures (chronic stress, acculturative stress, and ethnic discrimination) would be associated with higher insomnia symptoms and higher sleep variability, and lower sleep duration and lower sleep efficiency. We also hypothesized that the association of sociocultural stress severity and psychosocial stress burden with insomnia and objectively measured sleep dimensions would be stronger among unemployed versus employed Hispanic/Latino adults.

METHODS

Data source and sample

The HCHS/SOL is a community-based prospective cohort study of 16 415 self-identified Hispanic/Latino adults recruited from randomly selected households in four US field centers (Chicago, IL; Miami, FL; Bronx, NY; San Diego, CA) with baseline examination occurring between 2008 and 2011. The HCHS/SOL study sample and design have been described elsewhere [33, 34]. The HCHS/SOL SCAS was conducted between February 2010 and June 2011 to examine sociocultural correlates of health for more details see [35] in a subsample of HCHS/SOL participants (n = 5313). Similarly, the Sueño ancillary study recruited a subsample of HCHS/SOL participants (n = 2252) across all four sites from 2010 to 2013, aged 18–64 years and free of severe sleep-disordered breathing (apnea-hypopnea index [AHI] < 50/hour, no positive airway pressure treatment for sleep apnea) and narcolepsy, to undergo more detailed sleep assessment [27]. The mean time between HCHS/SOL SCAS and Sueño assessment was M = 663.9 days (SD = 160.5). Approval for the studies was obtained from institutional review boards at all institutions involved, and all participants provided informed consent.

All Sueño participants attended a clinic visit, where they completed questionnaires in either English or Spanish about their sleep, employment, and work status, and wore a wrist actigraph to measure their sleep for 1 week. Each participant was provided with an Actiwatch Spectrum (Philips Respironics, Murrysville, PA) wrist actigraph to be worn on their nondominant wrist for 1 week. At least 5 days of valid data was required for inclusion in analysis. All data were analyzed in 30-second epochs.

Measures

Sociocultural stress severity

All of the sociocultural stress variables were measured at the HCHS/SOL SCAS exam. Acculturation stress was measured using the abbreviated 17-item Hispanic Stress Inventory (HSI) that measures distress/worry associated with interpersonal, economic, and immigration conflict accompanying the process of adaptation and integration into a new nonnative culture within the past 3 months [21, 36]. Each question asks about the presence of a specific conflict (No [0] or Yes [1]), and then the extent of distress/worry associated with each conflict on a 5-point scale, where 1 corresponds to “not at all worried/tense” and 5 corresponds to “extremely worried/tense.” Responses were summed to compute a total score. Cronbach’s alpha for the acculturative stress intrafamilial and extrafamilial subscales ranged from .71 to .72 for the English language measure, and .74–.85 for the Spanish language measure in the HCHS/SOL SCAS. Ethnic discrimination was measured with the 17-item Brief Perceived Ethnic Discrimination Questionnaire – Community Version (Brief PEDQ-CV) [37]. The Brief PEDQ-CV measures exposure to four different types of ethnic discrimination: exclusion/rejection, stigmatization, discrimination at work/school, and threat aggression. Each question is scored on a 5-point scale, where 1 corresponds to “never happened” and 5 corresponds to “happened very often.” A summary score of the four types of ethnic discrimination was generated. The Cronbach’s alpha was .91 and .87, for the English and Spanish measures respectively, in the HCHS/SOL SCAS.

Stress burden

Chronic moderate/severe stress was measured with an 8-item instrument that assesses exposure and extent of psychological stress that spans for at least 6 months or more across multiple life domains, including work, relationships, finances, and personal or family health problems [38]. A summary score was generated that refers to a count of total number of stressors with a minimum of 6 months duration and self-endorsed moderately stressful to very stressful distress.

Sleep

All sleep indicators were measured at the HCHS/SOL Sueño exam. Insomnia symptoms were measured using the Insomnia Severity Index (ISI), a 7-item instrument which assesses the nature, severity, and impact of insomnia. Scores were analyzed as continuous and dichotomous variables; scores greater than or equal to 15 have 86.1% sensitivity and 87.7% specificity for detecting insomnia cases in community samples [39]. Actigraphic measures of sleep. Sleep duration, sleep efficiency, and sleep variability were calculated from actigraphy (for more information on scoring of actigraphy studies in Sueño see [40]) and treated as continuous variables. Average nightly sleep duration was based on average duration of time scored as sleep between sleep onset and sleep offset across all days. Sleep efficiency was calculated as sleep duration divided by time between sleep onset and sleep offset averaged across all main sleep intervals. Sleep variability was operationalized as the SD of sleep duration across all days.

Employment status

Employment status was ascertained from participant’s responses to the questions about their employment and shift work status at the HCHS/SOL Sueño exam. Responses were transformed into a three-level categorical variable: unemployed (those who reported being unemployed/seeking work, retired/disabled, student, and homemaker), employed-shift work (those who reported being employed and working either the night shift, split shift, irregular shift/on call, or rotating shift), or employed non-shift work (those who reported being employed and working the day or afternoon shift).

Covariates

Socio-demographics were self-reported and collected at the baseline HCHS/SOL exam, including: age, sex (male, female), Hispanic/Latino background (Central American, Cuban, Dominican, Mexican, Puerto Rican, South American, more than one), annual household income (<$20 000 or ≥$20 000), nativity status (born in US mainland, immigrated ≥10 years, immigrated <10 years), and education (less than high school, high school or beyond). Medical Conditions were derived from the baseline HCHS/SOL examination. Hypertension was defined as a systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, or use of antihypertensive medication [41]. Diabetes mellitus was defined as a fasting plasma glucose ≥ 126 mg/dL, 2-hour postload plasma glucose ≥ 200 mg/dL, HbA1c ≥ 6.5%, or use of hypoglycemic medications [42]. Prevalent coronary heart disease was defined as a physician diagnosis of heart attack, angina, or a history of coronary artery bypass surgery, balloon angioplasty, or stent placement in one or more coronary arteries based on self-report, or evidence of history of a myocardial infarction based on electrocardiogram examination [43]. Obstructive lung disease was based on a forced expiratory volume in 1 second to forced vital capacity ratio less than 0.70 [44]. At baseline, participants underwent limited channel sleep apnea monitoring at home using the ARES Unicorder 5.2 (B-Alert, Carlsbad, CA) [45]. The use of this device and the process for centralized scoring in HCHS/SOL have been described elsewhere [46]. Respiratory events identified based on a 50% or greater reduction in airflow lasting 10 seconds or more and associated with a 3% oxyhemoglobin desaturation, were used to compute the AHI. Health Behaviors were ascertained at the baseline HCHS/SOL examination. Alcohol use (never, former, current) and cigarette use (never, former, current) were derived from self-report. Body mass index (BMI) was calculated from measured height and weight at the baseline HCHS/SOL examination. Mood was measured at the HCHS/SOL Sueño exam. Depressive symptoms were measured with the 10-item version of the Center for Epidemiologic Studies Depression scale (CES-D10), which asks respondents to rate how often in the past week they have experienced depressive symptoms on a 4-point scale (0–3) [47–49]. Anxiety symptoms were measured with the State-Trait Anxiety Inventory-Trait Score (STAI) [50].

Statistical analyses

Analyses were weighted to adjust for disproportionate selection probabilities into HCHS/SOL, SCAS, and Sueño and to adjust for bias due to differential nonresponse at the household and person levels [33, 34]. All analyses accounted for cluster sampling and the use of stratification in sample selection. Estimates are applicable to the study target population, defined as Hispanics/Latinos, ages 18–64 years, living in the four study communities. There were a total of 1288 HCHS/SOL participants who participated in both SCAS and Sueño. Of these, 1198 had both stress and actigraphy measures and 1192 had complete data for the three stress scores, actigraphy, ISI, and employment status.

Acculturation stress and ethnic discrimination measures were transformed to z-scores to facilitate comparisons across stressor types. To confirm that there were no extreme correlations among the primary independent variables (ethnic discrimination, acculturation stress, chronic moderate/severe stress) a formal test of multicollinearity was performed with the unweighted data. The variance inflation factor (VIF) for each of the predictors was well below 10 (range = 1.29–1.38); as such, the three stress exposures can be examined in the same model. Further, the unweighted Pearson’s correlations among the three stress exposures ranged between r = .26–.39.

We conducted survey linear regressions to evaluate the association of each stress variable with ISI score and the actigraphic measures of sleep in individual and aggregate models with three progressive levels of adjustment: model 1 adjusted for study site, age, sex, and Hispanic/Latino background group; model 2 additionally adjusted for nativity status, household income, education, and employment status; and model 3 additionally adjusted for BMI, hypertension, coronary heart disease, lung disease, diabetes mellitus, alcohol use, cigarette smoking, and AHI. Individual models evaluated the association of each stress variable separately with the sleep outcomes. The aggregate models included all stress exposures in the same model. Effect modification by employment status (unemployed vs. employed) was assessed with the inclusion of an employment status*stress severity or stress burden cross product in the adjusted models. Separate interaction terms were entered for each of the stress variables in each of the individual models. For ease of interpretation, we plotted any significant interactions using marginal means. For these analyses, employment status was treated as a two-level categorical variable (unemployed vs. employed [shift work and non-shift work]).

We also conducted four sensitivity analyses. First, we assessed for linearity in the effect of stress on sleep duration. Categorizing sleep into short (<6 hours), intermediate (6–8 hours), and long (≥8 hours) duration, we confirmed that the relationship between sleep duration and stress was linear. Second, beyond modeling ISI as a continuous measure, we also dichotomized using the clinical cutpoint of ISI ≥ 15. Third, we re-ran all of the analyses with an additional adjustment for depressive symptoms and anxiety symptoms in the final models. Finally, because prior research has shown that the effect of psychosocial factors such as worries on sleep problems is greater among women than men [51], and that US-born Latina/os tend to exhibit worse sleep profiles than foreign-born Latina/os [25, 27], we tested for effect modification of the impact of stress on sleep by sex and nativity by evaluating the significance of sex*stress, and nativity*stress interaction terms.

RESULTS

Table 1 shows the weighted age and gender standardized (to the 2010 US Census) mean and prevalence distribution of the sociodemographic, psychosocial, health risk, and sleep outcomes for the 1192 participants that contributed information for these analyses. A substantive percentage was estimated to be employed (46.4%), and low income (47.7%). A minority were estimated to be shift workers (24.0%), and US-born (27.7%). Overall, 14.9% were estimated to have insomnia symptoms above the clinical cutoff, average actigraphy-assessed sleep duration was 404.9 (standard error [SE] = 2.5) minutes or 6.75 hours, and average sleep efficiency was 88.2% (SE = 0.2).

Table 1.

Weighted age and gender standardized sociodemographic, psychosocial, health risk, and sleep outcomes: HCHS/SOL Sueño & Sociocultural Ancillary Studies, N = 1192

	Overall (N = 1192)
Sociodemographic factors	% or M	SE	Range
Household income
Less than or equal to $20,000	47.7%	2.3
Nativity status
Born in the United States	27.7%	2.1
Educational attainment
Less than high school	29.3%	1.9
Employment status
Currently employed	46.4%	2.1
Shift worker	24.0%	2.0
Health risk factors
Body mass index (kg/m2)	29.5	0.3
Hypertension	18.5%	1.2
Obstructive lung disease	8.9%	1.3
Diabetes	12.8%	1.1
Coronary heart disease	4.5%	0.6
Apnea-hypopnea index (events/hour)	4.0	0.2
Alcohol use
Heavy Drinking	4.6%	0.8
Cigarette use
Current smoker	22.1%	1.7
Depressive symptoms (CESD-10)	6.9	0.3	0–30
Anxiety symptoms (STAI-Trait)	17.2	0.2	0–30
Stressors
Stress severity
Ethnic discrimination	25.0	0.3	17–85
Acculturation stress	12.5	0.5	0–85
Stress burden
Chronic moderate/severe stress	1.3	0.1	0–8
Sleep outcomes
Insomnia Severity Index (ISI) score	6.9	0.2	0–28
Insomnia ISI ≥ 15	14.9%	1.4
Sleep duration (minutes)	404.9	2.5
Sleep efficiency (%)	88.2	0.2
Sleep variability (SD of sleep duration—minutes)	76.2	1.7

M = mean; SE = standard error. Range included only for the validated psychological scales. The estimated mean and prevalence follows that of the survey area and have an age and gender distribution corresponding to the 2010 US census. Hypertension defined as BP > 140/90 and antihypertensive medication use. Obstructive lung disease was based on a forced expiratory volume in 1 second to forced vital capacity ratio less than 0.70. Diabetes mellitus was defined as a fasting plasma glucose ≥ 126 mg/dL, 2-hour postload plasma glucose ≥ 200 mg/dL, HbA1c ≥ 6.5%, or use of hypoglycemic medications. Heavy drinking defined as 7 drinks/week for women or 14 drinks/week for men. Prevalent coronary heart disease was defined as a physician diagnosis of heart attack or history of coronary artery bypass surgery, balloon angioplasty, or stent placement in one or more coronary arteries based on self-report. CESD-10 refers to the 10-item version of the Center for Epidemiologic Studies Depression scale. STAI-Trait refers to the State-Trait Anxiety Inventory-Trait Score. ISI refers to the Insomnia Severity Index.

Sociocultural stress severity, stress burden, and insomnia symptoms

Both acute sociocultural stress severity and chronic psychosocial stress burden had a positive association with ISI score across all individual models. Specifically, one SD increases in acculturation stress severity, ethnic discrimination stress severity, and chronic stress burden were each associated with 1.45 (SE = 0.25), 0.96 (SE = 0.26), and 1.27 (SE = 0.16) unit increases in the ISI score, respectively, in models that accounted for study site, age, sex, and Hispanic/Latino background group (ps < .001; Table 2). These estimates were slightly attenuated after further adjustment for nativity status, income, education, employment status, health risk behaviors, and medical conditions, but remained statistically significant. In aggregate models (all three stress variables entered in a single model), only acculturation stress severity and chronic stress burden remained positively and statistically associated with ISI score across all models (Table 2). Specifically, each 1 SD increase in the total number of moderate/severe chronic stress exposures, and 1 SD increase in the acculturation stress severity scale, were associated with a 1.04 (SE = 0.18, p < .001), and 0.75 (SE = 0.26, p < .01) unit increase in the ISI score holding constant the other two respective stressors. Ethnic discrimination severity was not associated with ISI score in the full models, though the estimates were in the same direction. While these trends were consistent in analyses with a dichotomized insomnia variable (Table 3), acculturation stress was no longer statistically associated with insomnia in the aggregate models.

Table 2.

Associations between sociocultural and psychosocial stressors^a and Insomnia Severity Index score: HCHS/SOL Sueño & Sociocultural Ancillary Study (N = 1192)

	Model 1b	Model 2c	Model 3d	Model 4e
	β (SE)	β (SE)	β (SE)	β (SE)
Individual modelsf
Acculturation stress, 1 SD	1.45 (0.25)***	1.50 (0.25)***	1.40 (0.24)***	0.50 (0.25)*
Ethnic discrimination	0.96 (0.26)***	0.86 (0.26)**	0.78 (0.27)**	0.27 (0.25)
Chronic stress burden	1.27 (0.16)***	1.23 (0.15)***	1.22 (0.16)***	0.71 (0.18)***
Aggregate modelsg
Acculturation stress	0.74 (0.26)**	0.83 (0.26)**	0.75 (0.26)**	0.20 (0.26)
Ethnic discrimination	0.29 (0.26)	0.22 (0.27)	0.20 (0.28)	0.08 (0.26)
Chronic stress burden	1.09 (0.18)***	1.03 (0.18)***	1.04 (0.18)***	0.68 (0.20)***

All models account for HCHS/SOL complex survey design and sampling weights.

^aAcculturation stress and ethnic discrimination variables were entered as z-scores.

^bModel 1 adjusts for study site, age, sex, and Hispanic/Latino background group.

^cModel 2 adjusts for Model 1+ nativity status, income, education, employment status.

^dModel 3 adjusts for Model 2 + body mass index, hypertension, coronary heart disease, lung disease, diabetes mellitus, apnea-hypopnea index, alcohol use, and cigarette smoking.

^eModel 4 adjusts for Model 3 + depressive symptoms and anxiety symptoms.

^fStressors tested in individual models.

^gStressors tested simultaneously in a single model.

*p < .05, **p < .01, ***p < .001.

Table 3.

Associations between sociocultural and psychosocial stressors^a and prevalence of insomnia (ISI ≥ 15): HCHS/SOL Sueño & Sociocultural Ancillary Study (N = 1192)

	Model 1b	Model 2c	Model 3d	Model 4e
	β (95% CI)	β (95% CI)	β (95% CI)	β (95% CI)
Individual modelsf
Acculturation stress, 1 SD	0.06 (0.04, 0.09)****	0.06 (0.03, 0.09)****	0.06 (0.03, 0.09)****	0.02 (−0.01, 0.04)
Ethnic discrimination, 1 SD	0.05 (0.02, 0.08)***	0.04 (0.01, 0.07)***	0.04 (0.01, 0.07)**	0.02 (−0.02, 0.04)
Chronic stress burden	0.06 (0.03, 0.08)****	0.05 (0.03, 0.07)****	0.05 (0.03, 0.07)****	0.03 (0.01, 0.05)***
Aggregate modelsg
Acculturation stress, 1 SD	0.03 (−0.00, 0.07)*	0.03 (−0.00, 0.07)*	0.03 (−0.01, 0.06)*	0.00 (−0.03, 0.03)
Ethnic discrimination, 1 SD	0.02 (−0.02, 0.05)	0.01 (−0.02, 0.05)	0.02 (−0.02, 0.05)	0.01 (−0.03, 0.04)
Chronic stress burden	0.05 (0.02, 0.07)****	0.05 (0.02, 0.07)****	0.04 (0.01, 0.07)****	0.03 (0.00, 0.05)**

CI, confidence interval. All models account for HCHS/SOL complex survey design and sampling weights.

^aAcculturation stress and ethnic discrimination variables were entered as z-scores

^bModel 1 adjusts for study site, age, sex, and Hispanic/Latino background group

^cModel 2 adjusts for Model 1+ nativity status, income, education, employment status,

^dModel 3 adjusts for Model 2 + body mass index, hypertension, coronary heart disease, lung disease, diabetes mellitus, alcohol use, and cigarette smoking

^eModel 4 adjusts for Model 3 + apnea-hypopnea index, depressive symptoms and anxiety symptoms

^fStressors tested in individual models.

^gStressors tested simultaneously in a single model.

*p < .10, **p < .05, ***p < .01, ****p < .001.

Sociocultural stress severity, stress burden, and objective markers of sleep

In contrast, most of the sociocultural stress severity and chronic psychosocial stress burden variables were not associated with any of the objective measurements made from 1-week actigraphy, namely, sleep duration, sleep efficiency, and sleep variability (Table 4). While the point estimates suggest mostly an inverse association between stress severity, stress burden, and objectively measured sleep, these estimates were not statistically significant, with the exception of the association of chronic stress burden and sleep duration. Each increase in the total count of chronic moderate/severe stress burden was associated with only a 3.5 (SE = 1.6, p < .05) minute increase in sleep duration in the fully adjusted aggregate model. Chronic stress burden was the only stress variable associated with an actigraphic measure of sleep.

Table 4.

Associations between sociocultural and psychosocial stressors and actigraphic sleep indicators: HCHS/SOL Sueño & Sociocultural Ancillary Study (N = 1192)

	Sleep duration (minutes)	Sleep efficiency (%)	Sleep variability (minutes)
	β (SE)	β (SE)	β (SE)
Individual modelsa
Acculturation stress, 1 SD	−3.33 (2.88)	0.12 (0.18)	−0.79 (1.76)
Ethnic discrimination, 1 SD	−4.46 (2.51)b	−0.05 (0.19)	−0.13 (1.75)
Chronic stress burden	2.21 (1.62)	−0.16 (0.15)	0.95 (1.10)
Aggregate modelsc
Acculturation stress, 1 SD	−3.47 (3.23)	0.27 (0.20)	−1.50 (1.91)
Ethnic discrimination, 1 SD	−4.17 (2.82)	−0.09 (0.21)	0.07 (1.83)
Chronic stress burden	3.46 (1.63)*	−0.20 (0.16)	1.25 (1.22)

All models account for HCHS/SOL complex survey design and sampling weights. Acculturation stress and ethnic discrimination variables were entered as z-scores. Model adjusts for site, age, sex, and Hispanic/Latino background, nativity status, income, education, employment status, body mass index, hypertension, coronary heart disease, lung disease, diabetes mellitus, apnea-hypopnea index, alcohol use, and cigarette smoking.

^a p = .07.

^bStress indicators tested in individual models.

^cStressors tested simultaneously in a single model.

*p < .05, **p < .001, ***p < .0001.

Employment status effect modification

Employment status modified the association of acculturation stress severity with insomnia symptom severity (p for interaction = .08). Each 1 SD increase in acculturation stress was associated with a 2.06 (SE = 0.34) increase in ISI among the unemployed versus a 1.01 (SE = 0.31) increase in ISI among the employed (Figure 1). No effect modification by employment status was observed for the actigraphy sleep outcomes.

Figure 1.

Effect of acculturation stress (in SD) on insomnia severity index score by employment status (solid line employed and dashed line unemployed) adjusted for age, sex, site, and ethnicity.

Sensitivity analyses

Although the estimates for the sociocultural stress severity and chronic stress burden variables associated with insomnia symptoms were markedly attenuated after further adjustment for the depressive symptoms and anxiety symptoms, the direction of the effects remained consistent (Table 2, Model 4). Acculturation stress severity (β = 0.50, SE = 0.25, p < .05), and chronic stress burden (β = 0.71, SE = 0.18, p < .001), remained positively associated with ISI score in the individual models, at statistically significant levels. In the aggregate models with final adjustment, only chronic stress burden (β =0.68, SE = 0.20, p < .001) continued to have a statistically significant association with ISI score. For the actigraphic sleep outcomes (Table 4), additional adjustment for depressive symptoms and anxiety symptoms slightly attenuated the point estimate of sleep duration associated with chronic stress burden (β = 3.3, SE = 1.8 minutes), though this was no longer statistically significant; none of the remaining stress exposures were associated with sleep efficiency or sleep variability. Additional exploratory analyses examining sex*sociocultural stress severity, or sex*stress burden, and nativity*sociocultural stress severity, or nativity*stress burden interactions with each of the sleep-related outcomes were not statistically significant at the p < .10 level.

Discussion

To our knowledge, this study is among the first empirical investigations on the association of disparate stress types (sociocultural stress severity and chronic stress burden) with various dimensions of subjectively and objectively measured sleep in a large and diverse population-based sample of Hispanics/Latinos in the United States. Overall, the results suggest that acculturation stress severity and total counts of chronic moderate/severe stressors have a consistent and positive association with an indicator of insomnia symptom severity, the ISI score, which remains significant with adjustment for potential confounders or mediators (e.g. anxiety, depressive symptoms). Employment status modified the association of acculturation stress severity with insomnia symptom severity, with evidence that associations were stronger in unemployed compared to employed individuals. Exposure to ethnic discrimination was not associated with any indicator of sleep in the adjusted aggregate models. Only chronic stress burden was associated with objectively measured sleep duration, but the magnitude of this observed relationship was small, of questionable clinical significance, and importantly was no longer statistically significant with further adjustment for depression symptoms and anxiety symptoms.

In general, the discrepant findings between stress and subjectively versus objectively measured sleep parallel previous research documenting inconsistencies in the relationship between stress and sleep as a function of sleep evaluation method (subjective vs. objective) [17], and study design (laboratory vs. ecological). Specifically, like prior research in non-Latina/o samples, we found that stress was more consistently associated with indicators of sleep satisfaction or sleep quality measured using self-report, and less so with objectively measured sleep duration, sleep efficiency, and sleep timing [17]. These findings support a multidimensional conceptualization of sleep health where the various dimensions of sleep are thought to reflect orthogonal but related constructs with subjective and, in most cases, objective analogs [7]. In conceptualizing sleep health as a multidimensional construct, one dimension of sleep, or one evaluation method of a specific sleep dimension is not prioritized as more “accurate” or “real”, or “precise” than another dimension. From this perspective, it is not presumed that an exposure would be similarly associated with each dimension of sleep health. As such, the observed statistically significant associations between stress and insomnia (an indicator of sleep satisfaction/quality), and the largely null associations between stress and actigraphy-based markers of sleep duration, sleep efficiency, and sleep timing may simply reflect the differential associations of stress with different dimensions of sleep health.

Consistent with previous research [24], acculturation stress severity and chronic stress burden had independent associations with insomnia symptom severity in the aggregated adjusted models that controlled for sociodemographic factors and health behaviors, while ethnic discrimination was not associated with any sleep outcome. These results underscore the need for multidimensional and comprehensive assessments of stress in research on stress processes and sleep, including the measurement of acuity/chronicity and the type of stressor. Indeed, both the measure of chronic stress burden and acculturation stress inquired about the presence of a given stressor, and also the severity of the stressor or the extent of distress due to the stressor. In contrast, the measure of ethnic discrimination inquired about lifetime exposure to discriminatory experiences due to race/ethnicity, and not the distress due to the range of discriminatory experiences. As such, this may have weakened the potential to detect a significant association between lifetime exposure to ethnic discrimination and past-week sleep behavior. While this is among the first studies to examine the associations of ethnic discrimination with objectively measured sleep parameters exclusively in a large probability-based sample of Latina/os, the null findings are consistent with prior studies in multi-ethnic samples that did not observe a statistically significant association between ethnic discrimination and actigraphic measures of sleep; though we acknowledge that the findings are quite mixed in this research area, and that more consistent findings have been observed with self-reported measures [20]. In contrast, the null findings between ethnic discrimination and insomnia may reflect differences in how insomnia was measured in Sueño versus other studies. In this study, insomnia was measured using a validated clinical scale (ISI), whereas previous studies used 1–3 selected survey items instead of a complete and validated scale. Additionally, we are unable to assess the independent effects of skin color (a potential marker of ascribed race) on sleep in Latina/o communities, or the extent to which the effect of ethnic discrimination on sleep is modified by skin color or ascribed race. Prior research has found that racially identified Black Latina/os have worse health than racially identified White Latina/os in the United States, and that reports of at least one episode of ethnic discrimination is high across Hispanic/Latina/o heritage groups in HCHS/SOL ranging from 64.9% to 98% [52, 53]. Future work should assess whether the effect of perceived ethnic discrimination on sleep is greater among Black Latina/os compared to White Latina/os.

Importantly, additional adjustment for depression and anxiety symptoms resulted in a marked attenuation in the insomnia symptom severity estimates associated with acculturation stress and chronic stress burden. While the large reduction might be due to confounding, it may also be due to mediation, in that emotional distress may be a likely pathway through which stress disrupts sleep. Future research with longitudinal designs should explore the causal relationships between stress severity and burden, insomnia symptom severity, and depression and anxiety symptoms in Hispanics/Latinos.

Employment status modified the effect of acculturation stress on insomnia symptom severity such that the insomnia estimates associated with acculturation stress were markedly higher among unemployed Hispanic/Latino adults compared to their employed counterparts. It is possible that adults experiencing heightened acculturation stress—defined as stress stemming from economic, linguistic, familial, and occupational conflict related to the process of adapting and integrating to the US culture—may be more vigilant about themselves, their families, and their surroundings, which in turn may serve as a precipitating factor for insomnia that negatively impacts the initiation and maintenance of sleep, creating a vicious cycle of vigilance and impaired sleep. Indeed, it is possible that the measure of acculturation stress is capturing experiences of marginalization and structural disadvantage in the 21st century, such as distresses related to immigration or documentation status, financial strain, or occupational stress that are precipitants of poor sleep. This acculturation stress-vigilance-impaired sleep cycle may be exacerbated among those who are unemployed compared to those who are employed, because unemployment may facilitate maladaptive behaviors such as excessive time in bed, daytime napping, and rumination that perpetuate insomnia, in individuals without access to healthcare and treatment, two predisposing factors to poor health.

While none of the sociocultural stress severity or chronic stress burden exposures were determinants of sleep efficiency or sleep variability, an increase in total moderate/severe chronic stress burden counts was associated with a very small increase in total sleep duration. Unlike our previous findings with the HCHS/SOL SCAS Latino/Hispanic cohort where we found a strong association between sociocultural stress severity measures and self-reported sleep duration [28], we did not find a consistent association between stress and actigraphic sleep measures. This may be due to differences in sample size or time of assessment. Alternatively, these results potentially suggest that the effect of stress on sleep duration may be overestimated when self-reported measures of sleep duration are used. The differential results may also suggest that self-reported measures of sleep duration actually measure sleep satisfaction or sleep quality rather than actual sleep duration.

Although the current results advance the scholarship on psychosocial determinants of sleep among Hispanics/Latinos, there are several limitations that warrant discussion. First, the data are cross-sectional, as such, we are unable to assess causality and directionality. Future studies using longitudinal data would allow researchers to evaluate whether insomnia symptoms increase stress, or vice-versa, or whether the relationships are bidirectional. Second, we do not have information on participants’ documentation status, occupation characteristics, or neighborhood-level factors, which may be third variables accounting for the observed relationships. Third, we had small sample sizes in select cells, which may have resulted in our being underpowered to detect significant differences, particularly regarding effect modification. Further, the HCHS/SOL sample was recruited from four large urban cities, as such the findings have limited generalizability to rural samples.

CONCLUSION

While sleep is a national public health priority [1], Hispanic/Latino sleep health is understudied. In a diverse sample of 1192 Hispanic/Latinos from four US cities, our findings suggest that acculturation stress severity in addition to chronic psychosocial stress burden are important and consistent correlates of insomnia symptom severity; and less consistent correlates of objectively measured markers of sleep derived from actigraphy. Future research should attempt to replicate these findings, and explore whether adding an adjunctive component that targets the resolution of acculturation stress to the gold standard nonpharmacological behavioral intervention for insomnia (namely, cognitive behavioral therapy for insomnia) improves insomnia among Hispanic/Latinos. Importantly, our results also suggest that unemployment status may exacerbate the risk of acculturation stress on insomnia in Hispanic/Latinos. Future research is needed to determine whether interventions that target early screening or detection of acculturation stress in unemployed Hispanics/Latinos would yield substantial public health benefit. Accordingly, population-level health efforts to improve sleep must sufficiently address upstream factors such as employment status, which appear to alter the stress to poor sleep disease course. To this end, multilevel interventions that provide employment assistance workshops to Hispanics/Latinos who are unemployed or who are at risk of a negative employment transition [54] may also confer additional health benefits by curbing the adverse effects of acculturation stress on risk of insomnia.

Funding

The Hispanic Community Health Study/Study of Latinos was carried out as a collaborative study supported by contracts from the National Heart, Lung, and Blood Institute (NHLBI) to the University of North Carolina (N01-HC65233), University of Miami (N01-HC65234), Albert Einstein College of Medicine (N01-HC65235), Northwestern University (N01-HC65236), and San Diego State University (N01-HC65237). The following Institutes/Centers/Offices contribute to the HCHS/SOL through a transfer of funds to the NHLBI: National Center on Minority Health and Health Disparities, the National Institute on Deafness and Other Communications Disorders, the National Institute of Dental and Craniofacial Research, the National Institute of Diabetes and Digestive and Kidney Diseases, the National Institute of Neurological Disorders and Stroke, and the Office of Dietary Supplements. The HCHS/SOL SCAS was funded by NHLBI RC2HL101649; The HCHS/SOL Sueño was funded by HL098297. C.A. is supported by K23 HL125748 and S.R.P. is supported by HL127307 from the National Heart, Lung, and Blood Institute, and M.D.C. is supported by K01 MH103511 from the National Institute of Mental Health, of the National Institutes of Health.

Conflict of interest statement. None declared.