Sleep Health Disparities in Children

Sleep Health Disparities in Children
Benjamin L. Sigurdson, Chaaya Nakra, Chesney Sickles-Jarvis, Chun T Au, Indra Narang
The Hospital for Sick Children

Abstract

Sleep is a critical determinant of child and adolescent health; however, evidence suggests that sleep outcomes are not distributed equally across populations and are influenced by multiple social and environmental determinants. This review identified three studies examining associations between sleep outcomes and socioeconomic status (SES), area-level income inequality, geographic location, race/ethnicity, and sex/gender. Across studies, lower individual- and household-level SES was consistently associated with poorer sleep outcomes, including shorter sleep duration, lower sleep quality, reduced adherence to sleep guidelines, and greater obstructive sleep apnea severity. Area-level income inequality was also associated with worsening sleep guideline adherence over time, independent of area wealth. Limited evidence further suggested disparities by race/ethnicity and gender, with Black and gender-diverse adolescents experiencing the poorest sleep outcomes. Geographic differences were comparatively understudied, with some evidence indicating poorer sleep-disordered breathing outcomes among urban youth. Collectively, findings suggest that social and structural disadvantage contributes to sleep health inequities among Canadian youth, while highlighting important gaps in the literature.

Overview

Three studies met inclusion criteria, all conducted in Canada. Two used cross-sectional designs (Park et al., 2022; Mitchell et al., 2024) and one used a three-wave longitudinal cohort design (Hunter et al., 2025). Sample sizes and populations varied substantially: Park et al. drew on a clinical sample of 188 children aged 4–17 years referred to two hospital-based sleep clinics for suspected obstructive sleep apnea (OSA), with OSA variables assessed objectively via polysomnography (PSG); Mitchell et al. (52,138 participants), and Hunter et al. (9,299 participants) both drew on the population-based COMPASS school survey (adolescents aged 12–19), with sleep assessed via self-report, and Hunter et al. additionally linking COMPASS data to 2016 Canadian Census data at the census division (CD) level. The study by Mitchell et al. was conducted during the first full year following COVID-19. Together, the three studies examine five distinct social/environmental exposures: household- and individual-level socioeconomic status (SES), area-level income inequality, geographic location/urbanicity, race and ethnicity, and sex/gender.

1. Socioeconomic Status

SES was examined in all three studies but operationalized in meaningfully different ways: as individual/household-level indicators (Park et al.; Mitchell et al.) vs. as area-level relative income inequality (Hunter et al.). These reflect distinct theoretical constructs (absolute material/social disadvantage vs. relative structural inequality), so findings are presented separately below before being synthesized.

1.1 Household- and Individual-Level SES

Park et al. assessed SES along five dimensions in their clinical sample: maternal education, father's income, occupational social class, geographic location (addressed separately in Section 2), and perceived (subjective) SES. After adjusting for age, sex, and BMI z-score, maternal education showed the strongest and most consistent association with OSA outcomes: children whose mothers had less than a high school education had nearly three times the odds of presenting with an apnea index (AI) ≥ 1 event/hour compared to children of university-educated mothers (OR = 2.96, 95% CI 1.05, 8.37). This was the only individually statistically significant exposure-outcome contrast reported across the entire study. Associations were directionally consistent, with lower maternal education being associated with more apnea events, higher apnea-hypopnea index (AHI) severity, and higher odds of moderate/severe apnea across all four outcome measures, though the remaining contrasts did not reach statistical significance. Father's income and occupational social class showed weaker and less consistent associations, with confidence intervals crossing the null for every outcome and, in the case of social class, direction of effect varying across outcomes (e.g., lower social class was associated with more apnea events but paradoxically lower odds of moderate/severe apnea). Perceived/subjective SES (based on the MacArthur Scale of Subjective Social Status) showed minimal, near-null associations with all four outcomes.

Mitchell et al., using a much larger population-based sample (N=52,138), measured SES via a composite score derived from four adolescent-reported items (going to bed hungry, relative financial comfort, having one's own bedroom, and COVID-19-related financial worry), scored on a 5-point scale. This study found a clearer, more precisely estimated monotonic gradient than Park et al.: adolescents in the lowest SES category slept approximately 28.6 fewer minutes per day on weekdays (95% CI -39.5, -17.6) and 17.5 fewer minutes per day on weekends (95% CI -31.2, -3.8) than those in the highest SES category, and had substantially lower odds of meeting sleep duration guidelines (AOR = 0.54, 95% CI 0.41, 0.70) and of reporting good sleep quality (AOR = 0.33, 95% CI 0.26, 0.43), compared to the highest SES category. Some evidence of a plateau at the upper end of the SES distribution was noted, with smaller and often non-significant differences between the two highest categories.

Despite differing substantially in sample size, population (clinical vs. population-based), SES measurement approach, and outcome ascertainment (objective PSG data vs. self-report), both studies point in a broadly consistent direction: lower individual/household-level SES, is associated with worse sleep outcomes in Canadian children and adolescents. The association was most robust and precisely estimated for sleep duration and guideline adherence; associations with objectively measured OSA severity/frequency (Park et al.) were directionally consistent but less statistically robust, plausibly reflecting the much smaller clinical sample. It is worth noting that "maternal education" and Mitchell's material-deprivation-style composite score are not identical constructs, even though both are commonly grouped under "SES."

1.2 Area-Level Income Inequality

Hunter et al. took a distinct, ecological approach, examining area-level income inequality (Gini coefficient, calculated at the census division level from 2016 Canadian Census after-tax household income) rather than individual- or household-level SES. Using longitudinal COMPASS data (n=9,299) across three annual waves (2016–17 to 2018–19), the study modelled a Gini*time interaction to test whether trajectories of movement guideline adherence diverged over time by baseline area-level inequality. Models adjusted for covariates at all three levels of the data structure: student-level (age, gender, racial/ethnic identity, weekly spending money), school-level (public/private status, urbanicity, school area-level income), and CD-level (population, median after-tax household income). CD median income was included as a separate covariate from the Gini coefficient itself, allowing the model to distinguish the effect of relative inequality from absolute area wealth. Adolescents in census divisions with greater income inequality showed significantly steeper declines over time in the odds of meeting sleep duration guidelines (OR = 0.83, 95% CI 0.74, 0.92 by 2017–18; OR = 0.77, 95% CI 0.70, 0.86 by 2018–19; p < 0.0001), and a similar pattern for combined sleep and physical activity guideline adherence (OR = 0.82, 95% CI 0.71, 0.94 by 2018–19; p = 0.0200). No significant time interaction was found for physical activity or screen time guidelines considered independently, and no heterogeneity by gender was detected.

This finding adds a structural dimension distinct from the individual-level SES findings above. Because CD-level median income was already controlled for, the result suggests that relative inequality within an area, not merely its absolute wealth, is associated with a specific and worsening reduction of adolescent sleep guideline adherence over time, even though this study did not find equivalent effects for physical activity or screen time in isolation.

2. Geographic Location/Urbanicity

Only Park et al. examined geographic location as a distinct social/environmental exposure, classifying children's residence (via postal code) as urban, periurban, or rural/remote in British Columbia and Quebec. Children in periurban and rural areas consistently showed fewer apnea events, lower AHI severity, and lower odds of OSA (both AI ≥ 1 event/h and moderate/severe apnea) relative to urban-dwelling children, although none of these associations reached statistical significance in the primary adjusted model. The authors note that the periurban-urban contrast became statistically significant for apnea severity in a sensitivity analysis additionally adjusting for ethnic background and craniofacial syndrome status.

The authors offer two tentative, untested explanations: their periurban sample happened to skew toward higher-SES households, and urban environments may carry greater exposure to air and light pollution, both plausible risk factors for OSA. It is worth noting that this is a single-study finding with wide, non-significant confidence intervals, and neither Mitchell et al. nor Hunter et al. examined urbanicity/geography as a primary exposure. Geographic/environmental determinants of youth sleep in Canada appear substantially understudied relative to individual-level SES.

3. Race and Ethnicity

Only Mitchell et al. examined race and ethnicity in relation to sleep health. Relative to White adolescents, all other racial/ethnic groups reported shorter weekday and average daily sleep duration (mean differences of approximately 5–15 min/day) and lower odds of meeting sleep guidelines, with Black adolescents showing the largest deficits (weekday duration: -14.3 min/day, 95% CI -21.6, -7.0; guideline adherence AOR = 0.61, 95% CI 0.52, 0.73, vs. White). These disparities were concentrated on weekdays; weekend sleep duration did not differ significantly by race/ethnicity for any group, suggesting racial/ethnic sleep gaps may be partly compensated for outside the school week. Sleep quality showed few significant racial/ethnic differences overall; the only significant contrasts were between the "Another" (multiracial/other) category and both Asian and White adolescents.

It is important to note that with only one study addressing this determinant, findings cannot yet be corroborated across the evidence base. Additionally, none of the three studies examined Indigenous identity as a distinct category (Mitchell et al.'s race/ethnicity categories are limited to Asian, Black, Latin, White, and Another), which is a notable gap given the well-documented health inequities affecting Indigenous children.

4. Sex and Gender

Mitchell et al. was also the only study to examine sex/gender differences, distinguishing female, male, and a combined "other" category (adolescents who described their gender differently or preferred not to disclose their sex/gender, combined due to small cell sizes). Differences between female and male adolescents were modest: females slept slightly less than males on weekdays (-1.7 min/day) but more on weekends (7.1 min/day), with no significant difference in average daily sleep duration, suggesting weekend "catch-up" sleep. Despite this, females were less likely than males to report good sleep quality (AOR = 0.57, 95% CI 0.54, 0.60).

By contrast, gender-diverse adolescents ("other") showed substantially worse outcomes across every measure: they slept approximately 15–30 fewer minutes per day than either female or male peers, on weekdays, weekends, and on average (weekday: -25.9 min/day vs. female, -27.6 min/day vs. male; weekend: -20.9 vs. female, -13.8 vs. male; average: -23.3 vs. female; -22.6 vs. male), and had markedly lower odds of meeting sleep guidelines and of reporting good sleep quality relative to both other groups (average guideline adherence AOR = 0.54 vs. females, 0.57 vs. males; sleep quality AOR = 0.60 vs. females, 0.34 vs. males).

Importantly, this is another single-study finding; however, the gender-diverse disparity in particular is large in magnitude and consistent across every outcome measured, making it a priority area for replication, especially since the "other" category collapses multiple distinct gender identities.

Summary and Synthesis

Across the three included studies, lower socioeconomic position emerged as the most consistently identified correlate of poorer sleep health. Evidence from both clinical and population-based samples suggested that children and adolescents experiencing socioeconomic disadvantage had poorer sleep outcomes, while longitudinal evidence indicates that adolescents living in areas with greater income inequality experienced worsening sleep guideline adherence over time, independent of area-level income. Together, these findings suggest that both individual-level disadvantage and broader structural inequalities may contribute to sleep health disparities.

Evidence regarding other social determinants was more limited but pointed to important inequities. Black adolescents and other racialized groups generally reported shorter sleep duration and lower sleep guideline adherence than White adolescents, while gender-diverse adolescents demonstrated consistently poorer sleep outcomes across all measures examined. Geographic disparities were less clear, although urban residence showed a tendency toward greater OSA burden than periurban or rural residence. Overall, the available evidence suggests that sleep health disparities among Canadian youth are socially patterned; however, the literature remains sparse, with important gaps related to Indigenous populations, environmental determinants, and the use of objective sleep measures.

References

1. Park JW, Hamoda MM, Almeida FR, et al. Socioeconomic inequalities in pediatric obstructive sleep apnea. Journal of Clinical Sleep Medicine. 2022;18(2):637-645. doi:10.5664/jcsm.9494
2. Mitchell J, Magier MJ, Duncan MJ, et al. Inequities in sleep duration and quality among adolescents in Canada. BMC Public Health. 2024;24(1). doi:10.1186/s12889-024-19974-w
3. Hunter S, Perala Z, Patte K, et al. Income inequality and adherence to 24-hour movement guideline recommendations among adolescents: a multilevel growth curve analysis using longitudinal data from three waves of the Cannabis, Obesity, Mental health, Physical activity, Sedentary behaviour and Smoking (COMPASS) study (2016–2019). Journal of Epidemiology and Community Health. Published online February 18, 2025:jech-2024-223176. doi:10.1136/jech-2024-223176