Abstract
Aim:
To examine the interaction between sleep and social determinants of health (SDOH) [race/ethnicity and socioeconomic status (SES)] on overweight/obesity in adolescents.
Design:
Cross-sectional.
Methods:
We conducted a secondary analysis using the 2017–2018 National Survey of Children’s Health data. We included adolescents (10–17 years old) who had sleep and body mass index (BMI) data available (n = 24,337) in analyses (samples with BMI <5th percentile excluded). Parents reported children’s sleep duration and regularity. High BMI (≥85th percentile) for age defines overweight/obesity. We selected SDOH (race/ethnicity, family income, primary caregiver education and neighbourhood condition) and covariates (age, sex, smoking, exercise and depression) using a hierarchical model-building approach. Accounting for complex survey design, logistic regression estimated the interaction between sleep and SDOH.
Results:
There were significant interactions between sleep duration and SDOH. The association between increasing sleep and decreasing odds of overweight/obesity only showed in the following subgroups: White, family income ≥400% federal poverty level (FPL) or primary caregiver’ education ≥ high school. Compared with these subgroups, Hispanic adolescents and adolescents whose family income was below 100% FPL and whose caregiver education was below high school had weakened and reversed associations. Sleep regularity was not associated with overweight/obesity.
Conclusions:
Increasing sleep duration was associated with a decreased risk of overweight/obesity, but the association was not present in adolescents from racial/ethnic minority groups (i.e. Hispanic) and those with low SES.
Impact:
The study findings suggest that associations between sleep and overweight/obesity vary by race and SES. Identification of additional mechanisms for obesity is needed for racial/ethnic minority groups and those from families with low SES. Also, the complexity of these relationships underscores the importance of community-based needs assessment in the design of targeted and meaningful interventions to address complex health conditions such as poor sleep and obesity.
Keywords: adolescent health, health disparity, obesity, public health nursing, sleep, social determinants of health
1 |. INTRODUCTION
Unhealthy sleep and overweight/obesity are co-existing epidemics in adolescents worldwide (Ogden et al., 2018). One in five children and adolescents are with overweight/obesity globally (WHO, 2021). Childhood obesity tracks into adulthood and contributes to 4.7 million premature deaths annually (DeBoer et al., 2015). Adolescents are disproportionately affected by unhealthy sleep compared with other age groups (Gariepy et al., 2020). Unhealthy sleep has been associated with weight indicators such as body mass index (BMI) in some (Breitenstein et al., 2019; Miller et al., 2018) but not all studies (Araújo et al., 2012). The inconsistent findings warrant a more nuanced examination of the association between sleep and weight status to elucidate individual differences in resiliency against or vulnerability to unhealthy sleep and weight gain.
Social determinants of health (SDOH) may act as moderators to potentiate or attenuate sleep-weight relationships (Bagley et al., 2015; Breitenstein et al., 2019). Children and adolescents who belong to racial/ethnic minority groups (e.g. Non-Hispanic Black and Hispanics) and those from families experiencing lower socioeconomic status (SES, e.g. low income and household education) are at risk for overweight/obesity-related health disparities (Ogden et al., 2018). In school-age children, low SES composite scores have been found to interact with unhealthy sleep (e.g. short sleep duration and poor sleep efficiency, variability in sleep onset time) to predict increased risk for overweight/obesity (Bagley et al., 2015; Breitenstein et al., 2019). However, there are very few studies focused on the interactive effects of contextual social determinants of health with sleep among adolescents.
Given the higher prevalence of unhealthy sleep and sensitive developmental stage during adolescence, identifying the subgroups of adolescents particularly vulnerable to sleep-related obesity risk is essential for prevention and intervention efforts to address health disparities.
2 |. BACKGROUND
Obesity reflects excess body fat or adipose tissue and overweight refers to excess weight in relation to height. BMI is a widely used indirect measure of adiposity to classify overweight and obesity status in population-based studies. Being overweight and obese during childhood and adolescence exponentially increases the risk of adulthood cardiometabolic diseases and subsequent morbidity and mortality (DeBoer et al., 2015). Unhealthy sleep, defined as short or long sleep duration, poor sleep quality, irregular timing of sleep and sleep disorders such as insomnia (Billings et al., 2021), has emerged as a non-traditional risk factor for overweight and obesity (Miller et al., 2018). The national guidelines for obesity prevention have underscored the importance of healthy lifestyle factors, including sufficient sleep, in children and adolescents (McGuire, 2012). However, the prevalence of overweight and obesity remains stagnant, warranting investigations of moderators that affect sleep-obesity associations.
Coinciding with high prevalence of overweight and obesity during adolescence, unhealthy sleep has been a public health concern for adolescents. Up to 68% of adolescents worldwide report at least one manifestation of unhealthy sleep (Gariepy et al., 2020). Unhealthy sleep may affect level of appetite-regulating hormones (i.e. leptin and ghrelin) and glucose metabolism (e.g. insulin secretion and insulin sensitivity), thereby leading to increased adiposity (Breitenstein et al., 2019). Recent meta-analyses support the positive associations between short sleep (below recommended age-appropriate hours) and concurrent and/or future overweight/obesity in children and adolescents (Miller et al., 2018). In addition, sleep duration, sleep efficiency (percentage of sleep time achieved between bed and wake times) and regularity in sleep onset timing have been negatively associated with BMI scores among children and adolescents (Bagley et al., 2015; Breitenstein et al., 2019; Chehal et al., 2022). Although the associations between sleep and weight status have been established, relations are not uniform across individuals (Araújo et al., 2012; Bagley et al., 2015; Breitenstein et al., 2019).
Theoretical and empirical evidence suggests the importance of examining social factors in relation to sleep and weight indicators. Based on the SDOH theoretical framework (Solar & Irwin, 2010), upstream, social determinants of health, such as socioeconomic status (i.e. family income and primary caregiver education), the built environment (i.e. neighbourhood conditions) and racial/ethnic groups, are considered major contributors to disparities in both sleep and overweight/obesity in adolescents (Jackson, 2017). Members of racial and ethnic minority groups are at disproportionate risk for unhealthy sleep (i.e. short sleep) and overweight/obesity (Billings et al., 2021; Jackson, 2017; Mead et al., 2022). Chronic psychological stress that accompanies cumulative experiences of discrimination may be a mechanism by which unhealthy sleep and overweight/obesity are more prevalent in such groups (Mead et al., 2022). Moreover, families of non-White race disparately reside in socioeconomically disadvantaged neighbourhoods that are burdened with low rates of education, high rates of poverty, unemployment and unsafe/unhealthy neighbourhood conditions (Billings et al., 2021), directly and indirectly jeopardizing sleep and healthy weight in adolescents.
Sleep does not appear to mediate the relationship between race, SES and health disparities in overweight/obesity among adolescents (Reither et al., 2014) and adults (Piccolo et al., 2013). However, the interaction between social risk factors and unhealthy sleep may amplify the risk of being overweight and obese during middle childhood (Bagley et al., 2015; Breitenstein et al., 2019). Compared with those with fewer social risks, children with unhealthy sleep (shorter duration and sleep irregularity) in conjunction with at least two cumulative social risk factors, including living in poverty, lower maternal education, high maternal stress and living in a single-parent household, were most likely to exhibit increases in BMI over time (Bagley et al., 2015). Consistent with dual-risk perspectives, unhealthy sleep may increase the allostatic load (i.e. the “wear and tear” on the physiological systems from dysregulated stress response and inability to maintain homeostasis) among racial/ethnic minorities and those of lower SES, thereby exponentiating the development of health conditions such as obesity (Bagley et al., 2015; Piccolo et al., 2013). Indeed, adolescents exposed to higher levels of social risk who are not receiving sufficient sleep (a physiological stressor) may have a greater burden of allostatic load which can lead to weight gain (Bagley et al., 2015).
There are several gaps in the literature. First, very few studies have involved adolescent populations who are at higher risk for unhealthy sleep. Adolescence serves as a sensitive period of development for behaviours and biological processes like sleep regulation and weight (Chehal et al., 2022). Exposures to these social factors during key developmental periods may set individuals on a trajectory for compromised or poor sleep and unhealthy weight status across the life course. Second, prior studies have focused on composite social risk scores. Although social risk factors (i.e. being a member of minority racial/ethnical group, family with low SES indicators) often co-exist, investigating their individual roles will serve as a foundation for understanding the most important social factors that change the sleep-overweight relationship. Third, prior studies have overwhelmingly utilized race/ethnicity as a covariate when assessing associations between child/adolescent sleep and weight indicators (Chehal et al., 2022). Short sleep duration and overweight/obesity are not equitably distributed across racial/ethnic groups, with racial/ethnic minorities more likely to experience social disadvantage secondary to present and historical experiences of racial/ethnic discrimination (Johnson et al., 2019). The contribution of race/ethnicity to sleep-related health disparities in adolescents, regardless of SES, will provide insight into additional socio-cultural dimensions of the relationship between sleep and overweight and obesity.
3 |. THE STUDY
3.1 |. Aim
The aim of this study was to examine the interaction between SDOH (i.e. race/ethnicity and SES indicators) and sleep duration/regularity on overweight/obesity in a nationally representative adolescent sample. We tested the following hypotheses: (1) the associations between decreasing sleep duration/regularity and overweight/obesity are stronger in racial/ethnic minorities, controlling for SES indicators; and (2) the associations are stronger among those with low SES indicators (e.g. primary caregiver education no more than high school education, family living with poverty, disadvantaged neighbourhoods), controlling for race/ethnicity.
3.2 |. Design
This study is a secondary analysis based on a cross-sectional dataset from the 2017 to 2018 National Survey of Children’s Health (NSCH).
3.3 |. Participants and data collection
The NSCH is a nationally representative survey on physical and psychosocial well-being of children aged 0–17 years, and the health care needs of their families and their communities. The survey was conducted online and via mail by the United States (US) Census Bureau. The US Census Bureau initially randomly selected addresses from non-institutionalized household across 50 states and the District of Columbia, stratified by state and households with children. Households were then screened to confirm eligibility, and one child per household was randomly selected as the participant in the survey. To obtain population-based estimates, each child for whom a survey was completed was assigned a weight that included adjustments for the base sampling weight, non-response, the selection of single child in a household, and demographic characteristics (US Census Bureau, 2019). The respondent was a parent or primary caregiver in the home who had the best knowledge of the child’s health (US Census Bureau, 2019).
The NSCH provides BMI data for adolescents 10–17 years of age and sleep data (duration and regularity) for participants between 4 months and 17 years. To achieve the study aims, adolescents (10–17 years old) who had complete data on sleep and BMI (n = 25,875) were included in this secondary analysis. Given the focus on overweight/obesity (vs. healthy BMI), we excluded participants with low BMI (<5th percentile), yielding 24,337 participants in final analyses.
3.4 |. Measures
3.4.1 |. BMI
The outcome variable in this study is BMI. Parents’ self-reported their child’s height and weight. BMI was calculated as weight (kg) divided by the square of height (m) and Center for Disease Control’s (CDC) BMI-for-Age Growth Charts were used to classify weight levels (US Census Bureau, 2019). According to the NSCH, overweight/obesity refers to BMI ≥85th percentile for age as defined by the CDC; and healthy BMI was defined as 5th percentile to less than 85th percentile for age (US Census Bureau, 2019).
3.4.2 |. Sleep variables
Our primary explanatory variable is sleep duration. Sleep duration was obtained by the primary caregiver’s answers to the question “During the past week, how many hours of sleep did this child get on most weeknights? (1 = less 6 h, 2 = 6 h, 3 = 7 h, 4 = 8 h, 5 = 9 h, 6 = 10 h and 7 = 11 or more hours)”. We used nighttime sleep duration as a continuous variable in adjusted models. For descriptive purposes, we also dichotomized sleep duration to determine whether children were getting age-appropriate recommended sleep hours. According to the American Academy of Sleep Medicine (AASM), 9–12 h of sleep per day on a consistent basis are recommended for 6–12 years old and 8–10 h for 13–18 years old to promote optimal health (Paruthi et al., 2016).
The secondary explanatory variable is sleep regularity. Sleep regularity was obtained by parent-report to the question “How often does this child go to bed at about the same time on weeknights?” Responses for this question were separated into five categories: 1 = always, 2 = usually, 3 = sometimes, 4 = rarely, 5 = never. Due to the small number of participants in categories 4 and 5, we merged response categories of “rarely” and “never” in data analyses.
3.4.3 |. Covariates and moderators
Based on the social-ecological model of sleep health (Grandner et al., 2015), SDOH theoretical framework (Solar & Irwin, 2010), and available variables in the dataset, we included the following variables as potential covariates: Individual level: birth history (preterm, low birth weight), depression, behavioural problems, cognitive problems, daily physical activity. Social level: race/ethnicity, household income, primary caregiver education, health insurance, neighbourhood factors (neighbourhood safety, neighbourhood support, neighbourhood conditions). Societal/environmental level: exposure to household mould, parent smoking and pesticide. Using a hierarchical model-building approach (see statistical analysis), we selected the following covariates for the final models: age, sex, depression, daily physical activity, race/ethnicity, household income, primary caregiver education, neighbourhood conditions and household smoking exposure.
Of the final selected covariates, SDOH were conceptualized as moderators for the sleep-BMI relationship. Specifically, a child’s Hispanic ethnic origin and race were based on the following categories: 1 = Hispanic, 2 = White, non-Hispanic, 3 = Black, non-Hispanic, 4 = Multi-racial (two or more race categories) or other (Asian, Indian/Alaska Native, Native Hawaiian/Other Pacific Islanders), non-Hispanic. Household income levels were determined based on the percentage of federal poverty level (FPL) of the household where the child lived. The categories for FPL were 1 = 0%–99%, 2 = 100%–299%, 3 = 300%–399%, and 4 = ≥ 400%, with an FPL of 0%–99% considered living in poverty. Education level was derived from the highest education of the primary caregivers. These categories were 1 = < high school, 2 = high school degree or GED, and 3 = > high school. In this study, categories 1 and 2 were merged due to the small number of primary caregivers with less than high school education. Neighbourhood condition was based upon the presence of detracting neighbourhood elements. These elements included (1) litter or garbage on the streets or sidewalk, (2) poorly kept or rundown housing or (3) vandalism such as broken windows or graffiti. The categories for this measure were 0 = neighbourhood does not have any detracting elements, 1 = neighbourhood has one detracting element, 2 = neighbourhood has two detracting elements and 3 = neighbourhood has three detracting elements.
Other covariates, such as depression (“Has a doctor or other health care provider EVER told you that this child has depression?”) and household smoking exposure (“Does anyone living in your household use cigarettes, cigars or pipe tobacco?”) were yes/no questions. Physical activity was obtained by parent-report to the question “During the past week, on how many days did this child exercise, play a sport or participate in physical activity for at least 60 minutes?” We categorized physical activity into “daily” versus “not daily”.
3.5 |. Validity and reliability
The NSCH dataset used in this study contains census-level data from all US states. Weights were assigned to samples to ensure the representativeness of the population. Although sleep duration and sleep regularity were measured using a single item, questions were asked in a way widely used in national health surveillance systems. In terms of weight indicator, BMI has been validated against more direct measures of body fat and is strongly correlated with metabolic outcomes (Wohlfahrt-Veje et al., 2014).
3.6 |. Statistical analysis
We performed data analyses with adjustment for complex survey design (cluster: household, stratum: state) and sampling weights using Stata 16.0 (svy package). We summarized sample characteristics using weighted means and percentages. At the bivariate level, we tested the difference in sleep and sample characteristics between adolescents with and without overweight/obesity using weighted Pearson Chi-square Statistic with the Rao–Scott correction (design-based F) for categorical variables and the Wald test for continuous variables (e.g. age). We used separate survey logistic regression models to estimate the associations between sleep duration/regularity (independent variables) and BMI (dependent variable), controlling for sex and age in the basic model. Additional covariates were selected using a hierarchical model-building approach. Individual, social, societal/environmental factors (see details in covariates) were selected and sequentially entered into models as a group. Models accounted for survey design and sampling weight.
To obtain a parsimonious model, covariates from each category were kept based on the Wald test for covariates’ significance (p < .05) and Archer Lemeshow goodness-of-fit statistic for model fit (p > .05). The final model included covariates at the individual (age, sex, depression, daily physical activity), social (race/ethnicity, family income, primary caregiver education and neighbourhood conditions) and societal/environmental (smoking) levels. To test the study hypotheses, we added interaction terms between each SDOH (race/ethnicity, income, primary caregiver education and neighbourhood conditions) and sleep variables to test the moderating effect of social risk factors. We tested the moderating models for sleep duration and sleep regularity separately. We used the Wald test and Archer–Lemeshow goodness-of-fit statistic to test the interaction terms and compare the model fit, respectively. Significant interactions were further examined and illustrated using Stata’s margins plots command. All test values were two-sided, with the significance level set at α = .05.
3.7 |. Ethical considerations
This study was deemed exempt by University of Delaware Institutional Review Boards.
4 |. RESULTS
4.1 |. Sample characteristics
The study participants (n = 24,337) ranged in age from 10 to 17 years, with a weighted mean age of 13.56 years old and linearized standard deviation of 0.03. Nearly half were female (49.35%) and White (51%). One-third of adolescents (n = 7181) had overweight/obesity. Whereas 86% of participants always/usually went to bed at the same time on weeknights, 37% (n = 8319) slept shorter than age-appropriate hours defined by the AASM.
4.2 |. Bivariate associations
Table 1 shows counts and weighted percentages/means of sample characteristics by BMI groups. Sleep shorter than age-appropriate hours (vs. healthy sleep duration) (design-based F = 24.21, p < .001) and rarely/never going to bed at the same time (vs. always/usually/sometimes) (design-based F = 8.94, p < .001) were significantly associated with overweight/obesity. Males had a slightly greater proportion of overweight/obesity than females (34.32% vs 31.66%, design-based F = 4.08, p = .04). Adolescents who had daily exercise (design-based F = 21.45, p < .001) for at least 60 min, or who were free of depression (design-based F = 22.86, p < .001) and household smoking exposure (design-based F = 27.87, p < .001) had a smaller proportion of overweight/obesity compared with their counterparts (p < .001). In terms of SDOH (Table 1), children who were non-White (vs. White) (design-based F = 23.11, p < .001) or whose families had income levels <400% FPL (vs. ≥400%) (design-based F = 38.86, p < .001), lived in a neighbourhood having at least one distracting element (vs none) (design-based F = 8.46, p < .001), and a primary caregiver who received ≤ a high school education (vs. > high school) (design-based F = 75.97, p < .001) also had a larger proportion of overweight/obesity.
TABLE 1.
Survey descriptive statisticsa of the study sample by BMI status (n = 24,337)
| BMI status, n (%) |
|||
|---|---|---|---|
| Overall sample | Healthy BMI (n = 17,156) | Overweight/obesityb (n = 7181) | |
| Age, M±SD | 13.56 (0.03) | 13.66±0.03 | 13.35±0.05* |
| Sex | |||
| Male | 12,553 (50.65) | 8491 (65.68) | 4062 (34.32) |
| Female | 11,784 (49.35) | 8665 (68.34) | 3119 (31.66) |
| Daily exercise | |||
| No | 20,051 (81.05) | 13,878 (65.33) | 6173 (34.67)*** |
| Yes | 4222 (18.95) | 3228 (74.17) | 994 (25.83) |
| Current depression | |||
| No | 22,530 (94.60) | 16,071 (67.64) | 6459 (32.36) |
| Yes | 1665 (5.40) | 992 (56.22) | 673 (43.78) |
| Parent education | |||
| >High school | 20,313 (67.04) | 14,780 (71.70) | 5533 (28.30)*** |
| ≤High school | 4024 (32.96) | 2376 (56.69) | 1648 (43.31) |
| Family income | |||
| 0%-99% FPL | 2728 (19.09) | 1656 (56.60) | 1072 (43.40)*** |
| 100%-299% FPL | 7491 (36.91) | 4933 (63.11) | 2558 (36.89) |
| 300%-399% FPL | 3542 (12.05) | 2502 (70.71) | 1040 (29.29) |
| ≥400% % FPL | 10,576 (39.14) | 8065 (76.43) | 2511 (23.57) |
| Race/Ethnicity | |||
| White, non-Hispanic | 17,185 (50.71) | 12,443 (72.40) | 4742 (27.60)*** |
| Hispanic | 2728 (25.72) | 1739 (60.33) | 989 (39.67) |
| Black, non-Hispanic | 1587 (14.08) | 930 (58.43) | 657 (41.57) |
| Multi-racial/other | 2837 (9.49) | 2044 (69.34) | 793 (30.66) |
| Neighbourhood condition | |||
| Good | 19,259 (75.62) | 13,783 (69.15) | 5476 (30.85)*** |
| 1 detracting element | 3069 (15.88) | 2089 (63.62) | 980 (36.38) |
| 2 detracting elements | 948 (4.84) | 621 (61.67) | 327 (38.33) |
| 3 detracting elements | 585 (3.67) | 351 (50.51) | 234 (49.49) |
| Household smoking | |||
| No | 20,357 (84.55) | 14,720 (68.59) | 5637 (31.41)*** |
| Yes | 3729 (15.45) | 2267 (59.31) | 1462 (40.69) |
| Sleep duration | |||
| Normal | 16,018 (62.71) | 11,598 (69.74) | 4420 (30.26)* |
| Insufficient | 8319 (37.29) | 5558 (62.50) | 2761 (37.50) |
| Sleep regularity | |||
| Always | 5880 (27.70) | 3987 (63.32) | 1893 (36.68)*** |
| Usually | 15,101 (57.83) | 10,933 (69.96) | 4168 (30.04) |
| Sometimes | 2396 (9.98) | 1624 (65.44) | 772 (34.56) |
| Rarely or never | 897 (4.49) | 565 (57.62) | 332 (42.38) |
Weighted percentage and weighted mean ± linearized standard deviation.
According to the Center for Disease Control's BMI-for-Age Growth Charts, overweight/obesity is defined as BMI ≥85th percentile and healthy BMI as 5–85th percentile for age. Chi-square (categorical variables) and Mann Whitney U (age) tests examined whether the overall differences in proportions of healthy BMI vs. overweight/obesity are different between groups with following significance indicators:
p < .05;
p < .01;
p < .001.
4.3 |. Adjusted associations between sleep and BMI
Table 2 shows results from logistic regression with adjustment for covariates. For sleep duration, every level increase in a nominal scale from 1 to 7 (representing the range from <6 to ≥11 h) was associated with a 7% decrease in the odds of overweight/obesity (OR = 0.93, 95% confidence interval [CI, 0.88, 0.99], p = .045). There was no association between sleep regularity and overweight/obesity after controlling for sleep duration and covariates. Hispanic (OR = 1.35, 95% CI [1.12, 1.62], p < .001) and Black, non-Hispanic (OR = 1.48, 95% CI [1.21, 1.81], p < .001) adolescents versus their White counterparts had 36% and 45% increased odds of overweight/obesity, respectively. Compared with family income ≥400% FPL, lower family income levels were associated with 23%–61% increased odds of overweight/obesity (p < .05), with the greatest odds shown in families whose income was <100% FPL (OR = 1.61, 95% CI [1.31, 1.98], p < .001). Adolescents whose primary caregiver had no more than a high school education (OR = 1.47, 95% CI [1.24, 1.75], p < .001) or lived in disadvantaged neighbourhoods with three detracting elements (OR = 1.64, 95% CI [1.14, 2.37], p = .02) had an increased odds of overweight/obesity. Being female, daily exercise and being free of depression and household smoking exposure remained significantly associated with decreased odds of having overweight/obesity in the adjusted models (p < .05).
TABLE 2.
| Model 1 | Model 2 | Model 3 | Model 4 | |
|---|---|---|---|---|
| Age | 0.92 (0.89, 0.95)*** | 0.92 (0.89, 0.95)*** | 0.92 (0.89, 0.95)*** | 0.92 (0.89, 0.95)*** |
| Female (vs. male) | 0.85 (0.74, 0.97)* | 0.85 (0.75, 0.97)* | 0.85 (0.75, 0.97)* | 0.85 (0.77, 0.97)* |
| Sleep duration | 0.93 (0.88, 0.99)* | 0.89 (0.83, 0.95)*** | 0.88 (0.82, 0.93)*** | 0.85 (0.78, 0.92)*** |
| Sleep regularity (vs. always) | ||||
| Usually | 1.03 (0.74, 1.43) | 1.03 (0.74, 1.43) | 1.03 (0.74, 1.44) | 1.01 (0.73, 1.43) |
| Sometimes | 0.79 (0.58, 1.09) | 0.80 (0.58, 1.09) | 0.79 (0.57, 1.09) | 0.78 (0.57, 1.07) |
| Rarely or never | 0.83 (0.58, 1.19) | 0.84 (0.59, 1.19) | 0.83 (0.59, 1.19) | 0.82 (0.58, 1.17) |
| Parent education ≤high school (vs. > high school) | 1.47 (1.24, 1.75)*** | 0.83 (0.48, 1.47) | 1.46 (1.23, 1.73)*** | 1.46 (1.23, 1.73)*** |
| Family Income (vs. ≥400% FPL) | ||||
| 0%-99% FPL | 1.61 (1.31, 1.98)*** | 1.59 (1.29, 1.95)*** | 1.60 (1.30, 1.97)*** | 0.75 (0.39, 1.42) |
| 100%-299% FPL | 1.45 (1.25, 1.67)*** | 1.43 (1.25, 1.66)*** | 1.45 (1.26, 1.67)*** | 0.91 (0.52, 1.61) |
| 300%-399% FPL | 1.23 (1.12, 1.49)* | 1.21 (1.01, 1.47)* | 1.22 (1.01, 1.48)* | 0.71 (0.31, 1.59) |
| Race/Ethnicity (vs. White) | ||||
| Hispanic | 1.35 (1.12, 1.62)** | 1.33 (1.11, 1.60)** | 0.62 (0.33, 1.18) | 1.34 (1.12, 1.61)** |
| Black, non-Hispanic | 1.48 (1.21, 1.81)*** | 1.48 (1.21, 1.81)*** | 1.15 (0.56, 2.37) | 1.48 (1.21, 1.81)*** |
| Multi-racial/other | 1.03 (0.85, 1.24) | 1.03 (0.85, 1.24) | 0.97 (0.51, 1.84) | 1.04 (0.86, 1.25) |
| Neighbourhood conditions (vs. no detracting element) | ||||
| 1 detracting element | 1.13 (0.92, 1.38) | 1.12 (0.92, 1.38) | 1.13 (0.93, 1.39) | 1.12 (0.92, 1.38) |
| 2 detracting elements | 1.05 (0.74, 1.51) | 1.05 (0.73, 1.50) | 1.05 (0.74, 1.50) | 1.05 (0.74, 1.50) |
| 3 detracting elements | 1.64 (1.14, 2.37)** | 1.64 (1.14, 2.37)** | 1.64 (1.14, 2.38)** | 1.64 (1.14, 2.36)** |
| Household smoking | 1.29 (1.09, 1.54)** | 1.29 (1.09, 1.54)** | 1.29 (1.08, 1.53)** | 1.31 (1.10, 1.55)** |
| Daily exercise (vs. no daily exercise) | 0.58 (0.48, 0.70)*** | 0.58 (0.48, 0.70)* | 0.58 (0.48, 0.70)* | 0.58 (0.48, 0.70) |
| Depression (vs. no depression) | 1.67 (1.34, 2.09)*** | 1.66 (1.33, 2.06)*** | 1.66 (1.33, 2.07)*** | 1.67 (1.33, 2.08)*** |
| Interaction terms | ||||
| Sleep duration × parent education ≤high school | 1.15 (1.01, 1.31)* | |||
| Sleep duration × race/ethnicity | ||||
| Hispanic | 1.20 (1.04, 1.40)* | |||
| Black, non-Hispanic | 1.05 (0.89, 1.26) | |||
| Multi-racial/other | 1.01 (0.87, 1.18) | |||
| Sleep duration×family income | ||||
| 0%-99% FPL | 1.20 (1.04, 1.40)* | |||
| 100%-299% FPL | 1.12 (0.98, 1.27) | |||
| 300%-399% FPL | 1.14 (0.94, 1.39) | |||
| Archer-Lemeshow test | F = 1.29, p = .23 | F = 1.12, p = .35 | F = 0.45, p = .91 | F = 0.64, p = .76 |
| Wald test | F = 11.35, p<.001 | F = 6.03, p<.001 | F = 7.66, p<.001 | |
Model 1: basic model accounting for complex sampling weights and covariates; Model 2: Model 1+ sleep duration and parent education interaction; Model 3: Model 1+ sleep duration and race/ethnicity interaction; Model 4: Model 1+ sleep duration and income interaction. The results presented were odds ratio (95% confidence interval). Interaction terms tested but not significant (not presented in table) include sleep duration and neighbourhood conditions, and sleep regularity and each SDOH variable. Sleep regularity and duration were tested separately.
According to the Center for Disease Control's BMI-for-Age Growth Charts, overweight/obesity is defined as BMI ≥85th percentile and healthy BMI as 5–85th percentile for age. Archer-Lemeshow test: p > .05 indicates good model fit. Wald test: p < .05 indicates adding the interaction term improves model fit.
p < .05;
p < .01; p < .001.
4.4 |. Moderating effect of SDOH
As shown in Table 2, race/ethnicity, family income, and primary caregiver education were significant moderators for sleep duration and overweight/obesity, such that the associations between increasing sleep duration and decreasing odds of overweight/obesity were significant only in the following social subgroups: White race (OR = 0.88, 95% CI [0.82, 0.95], p < .001), primary caregiver education >high school (OR = 0.89, 95% CI [0.83, 0.95], p < .001) and family income ≥400% FLP (OR = 0.85, 95% CI [0.78, 0.92], p < .001). Compared with White adolescents, Hispanic adolescents showed weakened and even reversed sleep-obesity association (OR = 1.20, 95% CI [1.04, 1.40], p = .01). Having family income <100% FPL (vs ≥400%) (OR = 1.20, 95% CI [1.04, 1.40], p = .02) and the primary caregiver having education ≤ high school (vs. > high school) (OR = 1.15, 95% CI [1.01, 1.31], p = .04) also attenuated/reversed the negative association between sleep duration and odds of having overweight/obesity (Table 2). More specifically, as shown in the margins plot (Figure 1), compared with non-White (i.e. Hispanic) and lower SES groups (i.e. family income <100% FPL and primary caregiver having education ≤ high school), adolescents with advantaged social factors (i.e. being White, family income ≥400% FPL and primary caregiver education > high school) had a smaller probability of overweight/obesity and the group differences in overweight/obesity probability became larger with increasing sleep duration (p < .05). The Archer–Lemeshow test suggests that models were a good fit for the data (p > .05) and adding the interaction terms showed a statistically significant improvement in model fit (Wald test, p < .05). Neighbourhood condition was not a significant moderator between sleep duration and overweight/obesity (p > .05). Also, there were no significant interactions between each SDOH variable (race/ethnicity, parent education, family income and neighbourhood condition) and sleep regularity (p > .05; data not presented in Table 2).
FIGURE 1.
Interaction between sleep duration and social determinants of health on overweight/obesity. The margins plots are derived using the Margins command in Stata based on logistic regression with complex survey design. The plots show the marginal effects of social determinants of health on probability of overweight/obesity (y-axis) with 95% confidence intervals when sleep duration (x-axis) is held constant at different values. *The associations between increasing sleep duration and decreasing probability of overweight/obesity were significant in subgroups: primary caregiver education ≥high school, family income ≥400% FLP and White (p < .05)
5 |. DISCUSSION
To the best of our knowledge, this is one of the first studies to examine how race/ethnicity and SES indicators amplify or attenuate the association between sleep and overweight/obesity using a nationally representative US sample of adolescents. Increasing sleep duration was associated with decreasing odds of overweight/obesity, with more robust association shown in White adolescents and those with family income≥400% FPL or primary caregiver > a high school education. These subgroups also had the lowest odds of overweight/obesity and the group differences increased with longer sleep duration, especially compared with Hispanic adolescents or those with family income <100% FPL and primary caregiver receiving no more than a high school education. Although neighbourhood condition was not a significant moderator, poor neighbourhoods with 3 detracting elements were independently associated with greater odds of having overweight/obesity. Sleep regularity was not independently associated with overweight/obesity or significantly interacted with SDOH on associations with overweight/obesity. Our findings inform future studies investigating the importance of sufficient sleep duration in conjunction with high SES to combating overweight/obesity.
More than one-third of our sample was classified as short sleepers or being overweight/obese. Consistent with the prior literature, overweight and obesity appear to differentially affect Hispanics, non-Hispanic Black and those with lower socioeconomic indicators in our sample (Ogden et al., 2018). Hispanic and non-Hispanic Black families and those with low SES disparately reside in disadvantaged neighbourhoods due to “redlining”, the historically discriminating US housing policy (Billings et al., 2021). Lack of access to healthy food options, green space and/or safe environments for physical activity may contribute to increased risk for overweight/obesity (Jackson, 2017). Similarly, elements that depict poor neighbourhood conditions such as vandalism/graffiti, rundown housing and cleanliness jeopardize healthy sleep duration in adolescents (Billings et al., 2021).
Adolescents with longer sleep duration on weekday nights were less likely to have overweight/obesity, which parallels prior research findings suggesting insufficient sleep as a risk factor for overweight and obesity (Miller et al., 2018; Morrissey et al., 2020). Proposed mechanisms underlying sleep-related overweight/obesity variations include metabolic changes affecting appetite-regulating hormone, low physical activity and increased food intake relative to energy expenditure (Breitenstein et al., 2019). Controlling for sleep duration, we did not find significant associations between sleep regularity and overweight/obesity in adolescents. Using the NSCH dataset in previous years, Chehal et al. (2022) found that some variability (vs. always regular) in weeknight bedtime is associated with lower odds of obesity, although there were no additional differences between those with extensive variability and always regular in bedtime. The discrepancy with our study may be attributed to different sets of covariates and classification of BMI levels (overweight/obesity vs. obesity). In contrast, lower sleep regularity measured using sleep diaries was associated with greater BMI in first-year college students (Wong et al., 2022). Mixed findings make this a continued gap and further research is needed to explore objectively measured sleep regularity with a longitudinal design.
Our findings of the interaction between sleep duration and SDOH both support and extend previous work. Obtaining sufficient sleep is associated with lower odds of obesity development among White adolescents and those with higher SES. Sufficient sleep may not be as protective against obesity risk in ethnic/minority and low SES groups, compared with their counterparts. These findings did not support our a priori hypothesis that the associations between shorter sleep and greater risk for overweight/obesity were more pronounced in adolescents in minority racial/ethnic and low SES groups. Prior research has also reported mixed results regarding the moderating role of race/ethnicity and SES (Bagley et al., 2015; Breitenstein et al., 2019). The associations between sleep duration and overweight/obesity did not vary by early-life SES calculated from caregiver education and household income among school-age children (Breitenstein et al., 2019) or ethnicity among low-income preschoolers (Vézina-Im et al., 2017). However, our findings, to some extent, align with study results suggesting that high SES early in life may buffer strong, negative associations between sleep duration and percent body fat and serve as a protective factor during middle childhood (Breitenstein et al., 2019). Additionally, whereas sleep extension intervention facilitates weight loss in adolescents under caloric restriction (Moreno-Frías et al., 2020), one experimental study showed that sleep extension intervention significantly improved sleep hours (by approximately 1 h) only among non-Hispanic whites but not adolescents of racial/ethnic minority status (Tavernier & Adam, 2017).
The lack of sleep-obesity association in racial/ethnic minority groups and those with low SES may be due to the multifaceted nature of obesity development in these groups. Compared with sleep impact, other risk factors such as poorer diet quality may explain a greater variance in overweight/obesity in those vulnerable groups. Since these vulnerable groups continue to live in socially disadvantaged environments, the cumulative and intergenerational differences in inopportune exposures could contribute to disparate risk of physiological dysregulation and its cascade of metabolic health consequences. As such, sleep duration on its own may not provide enough protection, and future obesity interventions may consider social and family-level changes in conjunction with multimodal health behaviour changes that are culturally relevant and place-based (because of pervasive racial residential segregation) (Jackson, 2017). Meanwhile, adolescents who have positive social risk factors and are having sufficient sleep may have a lower burden of allostatic load, thereby less likely to have overweight/obesity. Such that, increasing sleep duration enlarges the gap in overweight/obesity risk between socially disadvantaged groups (i.e. Hispanics, FPL < 100% and primary caregiver receiving < high school education) and those with advantaged social factors (i.e. being White and high SES status). Despite the nationwide “Hispanic paradox” (refers to the epidemiological findings of comparable/better health outcomes but lower SES than the non-Hispanic White) in obesity prevalence (Valencia et al., 2020), Hispanic children and adolescents are more vulnerable to obesity than their White peers (Ogden et al., 2018). The integration of multimodal health disparity pathways and sleep mechanisms is essential to understand and address obesity health disparities in adolescents with diverse racial/ethnic and socioeconomic backgrounds.
5.1 |. Implications for practice and future research
Sleep and obesity disparities disproportionately affect the same groups who suffer overall health disparities (Billings et al., 2021). Moreover, the negative impact of SDOH can accumulate over a lifetime, and intergenerationally transmit through family systems, and ultimately alter life course health trajectories (Piccolo et al., 2013; Solar & Irwin, 2010). These long-term public health impacts highlight the necessity for screening and intervention for unhealthy sleep and weight status early in the life course. Such tasks often fall onto paediatric providers, who are burdened with addressing multiple health domains during short well-child visits (Billings et al., 2021). Unhealthy sleep and overweight/obesity are each multi-etiological and complex in that they are determined by a myriad of factors at the individual, social and environmental levels. The addition of an interdisciplinary team of health care providers (e.g. school nurses, paediatric dentists and social workers), school teachers, religious leaders and other community stakeholders is needed to address sleep and cardiometabolic health inequities (Billings et al., 2021).
The current study suggests that the influence of socio-contextual factors may vary by race/ethnicity and SES. Prior to intervention development, more work is needed to better understand the complex mechanisms (beyond sleep behaviours) among these at-risk groups. With machine learning techniques, big data resources such as national surveillance, electronic health record and biobanks can be exploited to examine the interplay between sleep and other individuals (e.g. diet), family (e.g. parenting), social (e.g. social norms and cultural perception) and systemic factors and identify the most important factors for obesity development, particularly among racial/ethnic minority and low SES groups. Also, over a lifetime, this cyclical relationship of poor sleep leading to obesity and then obesity leading to worse sleep will only further health inequities (Jackson, 2017). Use of longitudinal design in future studies will help tease apart the likely bidirectional relationship between sleep duration and overweight/obesity in the context of SDOH. In terms of interventional studies, lack of acknowledgment of large-scale social factors will impede the success of individual health behaviour interventions to improve sleep and obesity, which will further exacerbate health inequities. As such, consideration of multi-level biobehavioural and socio-contextual factors, using youth, family and community-based perceptions are essential for the success of future obesity interventions, particularly those aimed at racial/ethnic minority groups (Billings et al., 2021).
5.2 |. Study limitations
Several potential limitations should be considered in the interpretation of results. First, the cross-sectional design does not support causal inferences. Reciprocal associations between sleep and overweight/obesity are likely and should be considered especially with longitudinal datasets to examine the trajectory of BMI in response to changes in sleep over time. Second, sleep and BMI measures relied on caregiver reports, which may pose a potential risk for bias. For example, parents tend to overestimate child sleep time, which may explain the lower rates of insufficient and irregular sleep in our sample, compared with a US national study using self-report (Wheaton et al., 2018). Additionally, using a single item to capture sleep duration and sleep irregularity limits construct variability and may impede its proper capture. Third, meta-analyses suggest that sleep more than age-appropriate hours predicts a greater risk of obesity in adults (Liu et al., 2019). However, there were only 1.54% of adolescents in this secondary analysis classified as long sleepers, thereby not supporting further examination of U-shaped associations between sleep and BMI. Fourth, due to data availability, we did not account for other behavioural and contextual differences related to weight status, especially dietary intake, screen time/screen-based sedentary behaviours. Access to calorie-dense/nutrition-poor foods, cultural habits for sleep and diet as well as screen behaviours could explain the observed moderating effect. Finally, there may be other moderators/mediators in the statistical model, such as sex, household smoking, daily exercise and depressive symptoms. Given the focus of this study, we did not test the moderation, mediation or moderated mediation of these variables. Future research is needed to identify the complicated relationships among multi-level variables.
6 |. CONCLUSION
Given the growing disparities in sleep health and obesity, our findings extend previous research by elucidating how the relationship between sleep and overweight/obesity varies based on race/ethnicity and SES among adolescents aged 10–17 years old. Longer sleep is associated with a lower overweight and obesity risk among adolescents who are White and from families with high SES, compared with their counterparts. SDOH have implications for the development of nurse-led weight management programs, as focusing on individual behaviours in the absence of consideration of social contributions may be unsuccessful. Future research is needed to investigate the most important contributing factors among adolescents from racial/ethnic minority groups and families with low SES to understand their vulnerability to overweight and obesity.
Footnotes
CONFLICT OF INTEREST
No conflict of interest has been declared by the authors. The data utilized in the submitted manuscript is publicly available from the NSCH website. There is a statistician on the author team (Ming Ji). Main statistical methods/approaches: descriptive, graphical methods, parametric & nonparametric tests and logistic regression. The authors agree to take responsibility for ensuring that the choice of statistical approach is appropriate and is conducted and interpreted correctly as a condition to submit to the journal.
PEER REVIEW
The peer review history for this article is available at https://publons.com/publon/10.1111/jan.15513.
NO PATIENT OR PUBLIC CONTRIBUTION
This is a secondary analysis.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available in Data Source Center for Child&Adolescent Health at https://urldefense.com/v3/__https://www.childhealthdata.org/help/dataset/2017-2018-combined-nsch-data-set-codebook-instruction__;!!N11eV2iwtfs!uw9SAt5WFrWN8MXSPGUI_b_psxMFlq5×1Kd1NAqJiTMfn05w5XTK-81fDRArTVw_5OZnzucMg9dW$.
REFERENCES
- Araújo J, Severo M, & Ramos E (2012). Sleep duration and adiposity during adolescence. Pediatrics, 130(5), e1146–e1154. 10.1542/peds.2011-1116 [DOI] [PubMed] [Google Scholar]
- Bagley EJ, Kelly RJ, & El-Sheikh M (2015). Longitudinal relations between children’s sleep and body mass index: The moderating role of socioeconomic risk. Sleep Health, 1(1), 44–49. 10.1016/j.sleh.2014.12.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Billings ME, Cohen RT, Baldwin CM, Johnson DA, Palen BN, Parthasarathy S, Patel SR, Russell M, Tapia IE, Williamson AA, & Sharma S (2021). Disparities in sleep health and potential intervention models: A focused review. Chest, 159(3), 1232–1240. 10.1016/j.chest.2020.09.249 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Breitenstein RS, Doane LD, & Lemery-Chalfant K (2019). Early life socioeconomic status moderates associations between objective sleep and weight-related indicators in middle childhood. Sleep Health, 5(5), 470–478. 10.1016/j.sleh.2019.04.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chehal PK, Shafer L, & Cunningham SA (2022). Examination of sleep and obesity in children and adolescents in the United States. American Journal of Health Promotion, 36(1), 46–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- DeBoer MD, Gurka MJ, Woo JG, & Morrison JA (2015). Severity of metabolic syndrome as a predictor of cardiovascular disease between childhood and adulthood: The Princeton Lipid Research Cohort Study. Journal of the American College of Cardiology, 66(6), 755–757. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gariepy G, Danna S, Gobiņa I, Rasmussen M, de Matos MG, Tynjälä J, Janssen I, Kalman M, Villeruša A, & Husarova D (2020). How are adolescents sleeping? Adolescent sleep patterns and sociodemographic differences in 24 European and north American countries. Journal of Adolescent Health, 66(6), S81–S88. [DOI] [PubMed] [Google Scholar]
- Grandner MA, Jackson NJ, Izci-Balserak B, Gallagher RA, Murray-Bachmann R, Williams NJ, Patel NP, & Jean-Louis G (2015). Social and behavioral determinants of perceived insufficient sleep. Frontiers in Neurology, 6, 112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jackson CL (2017). Determinants of racial/ethnic disparities in disordered sleep and obesity. Sleep Health, 3(5), 401–415. 10.1016/j.sleh.2017.08.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Johnson DA, Jackson CL, Williams NJ, & Alcántara C (2019). Are sleep patterns influenced by race/ethnicity – A marker of relative advantage or disadvantage? Evidence to date. Nature and Science of Sleep, 11, 79–95. 10.2147/NSS.S169312 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu W, Zhang R, Tan A, Ye B, Zhang X, Wang Y, Zou Y, Ma L, Chen G, Li R, & Moore JB (2019). Long sleep duration predicts a higher risk of obesity in adults: A meta-analysis of prospective cohort studies. Journal of Public Health (Oxf), 41(2), e158–e168. 10.1093/pubmed/fdy135 [DOI] [PubMed] [Google Scholar]
- McGuire S (2012). Institute of Medicine (IOM) early childhood obesity prevention policies. The National Academies Press. Advances in nutrition, 3(1), 56–57. 10.3945/an.111.001347 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mead MP, Vargas EA, & Knutson KL (2022). Racial disparities in sleep: Potential mediation by discrimination and psychological distress. Journal of Racial and Ethnic Health Disparities. Advance Online Publication. 10.1007/s40615-022-01247-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miller MA, Kruisbrink M, Wallace J, Ji C, & Cappuccio FP (2018). Sleep duration and incidence of obesity in infants, children, and adolescents: A systematic review and meta-analysis of prospective studies. Sleep, 41(4), zsy018. [DOI] [PubMed] [Google Scholar]
- Moreno-Frías C, Figueroa-Vega N, & Malacara JM (2020). Sleep extension increases the effect of caloric restriction over body weight and improves the chronic low-grade inflammation in adolescents with obesity. Journal of Adolescent Health, 66(5), 575–581. [DOI] [PubMed] [Google Scholar]
- Morrissey B, Taveras E, Allender S, & Strugnell C (2020). Sleep and obesity among children: A systematic review of multiple sleep dimensions. Pediatric Obesity, 15(4), e12619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ogden CL, Fryar CD, Hales CM, Carroll MD, Aoki Y, & Freedman DS (2018). Differences in obesity prevalence by demographics and urbanization in US children and adolescents, 2013–2016. JAMA, 319(23), 2410–2418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Paruthi S, Brooks LJ, D’Ambrosio C, Hall WA, Kotagal S, Lloyd RM, Malow BA, Maski K, Nichols C, & Quan SF (2016). Consensus statement of the American Academy of sleep medicine on the recommended amount of sleep for healthy children: Methodology and discussion. Journal of Clinical Sleep Medicine, 12(11), 1549–1561. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Piccolo RS, Yang M, Bliwise DL, Yaggi HK, & Araujo AB (2013). Racial and socioeconomic disparities in sleep and chronic disease: Results of a longitudinal investigation. Ethnicity & Disease, 23(4), 499–507. [PMC free article] [PubMed] [Google Scholar]
- Reither EN, Krueger PM, Hale L, Reiter EM, & Peppard PE (2014). Ethnic variation in the association between sleep and body mass among US adolescents. International Journal of Obesity, 38(7), 944–949. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Solar O, & Irwin A (2010). A conceptual framework for action on the social determinants of health. WHO Document Production Services. Retrieved March 18, 2022, https://www.who.int/publications/i/item/9789241500852
- Tavernier R, & Adam EK (2017). Text message intervention improves objective sleep hours among adolescents: The moderating role of race-ethnicity. Sleep Health, 3(1), 62–67. [DOI] [PubMed] [Google Scholar]
- US Census Bureau. (2019). 2018 National Survey of Children’s Health Methodology Report. Retrieved December 23, from https://www2.census.gov/programs-surveys/nsch/technical-documentation/methodology/2018-NSCH-Methodology-Report.pdf
- Valencia A, Zuma BZ, Spencer-Bonilla G, López L, Scheinker D, & Rodriguez F (2020). The Hispanic paradox in the prevalence of obesity at the county-level. Obesity Science & Practice, 7(1), 14–24. 10.1002/osp4.461 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vézina-Im LA, Hughes SO, Baranowski T, & Nicklas TA (2017). Association between sleep duration and body mass index among US low-income preschoolers. Obesity, 25(10), 1770–1775. [DOI] [PubMed] [Google Scholar]
- Wheaton AG, Jones SE, Cooper AC, & Croft JB (2018, Jan 26). Short sleep duration among middle school and high school students – United States, 2015. MMWR. Morbidity and Mortality Weekly Report, 67(3), 85–90. 10.15585/mmwr.mm6703a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- WHO. (2021). Fact sheet – Obesity and overweight. Retrieved March 15, from https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight
- Wohlfahrt-Veje C, Tinggaard J, Winther K, Mouritsen A, Hagen C, Mieritz M, de Renzy-Martin K, Boas M, Petersen J, & Main K (2014). Body fat throughout childhood in 2647 healthy Danish children: Agreement of BMI, waist circumference, skinfolds with dual X-ray absorptiometry. European Journal of Clinical Nutrition, 68(6), 664–670. [DOI] [PubMed] [Google Scholar]
- Wong PM, Barker D, Roane BM, Van Reen E, & Carskadon MA (2022). Sleep regularity and body mass index: Findings from a prospective study of first-year college students. Sleep Advances, 3(1), zpac004. 10.1093/sleepadvances/zpac004 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available in Data Source Center for Child&Adolescent Health at https://urldefense.com/v3/__https://www.childhealthdata.org/help/dataset/2017-2018-combined-nsch-data-set-codebook-instruction__;!!N11eV2iwtfs!uw9SAt5WFrWN8MXSPGUI_b_psxMFlq5×1Kd1NAqJiTMfn05w5XTK-81fDRArTVw_5OZnzucMg9dW$.