Abstract
Individual characteristics and social drivers of health (SDOH) likely interact to shape 24-hr movement behaviors (24HMB), yet most research has focused on single factors. We identified intersectional subgroups and examined disparities in meeting physical activity (PA), sedentary behavior (SB), sleep, and combined 24HMB recommendations. We performed a cross-sectional latent class analysis of 2022–2023 National Survey of Children’s Health data (n = 68,000, ages 6–17) using sex, race/ethnicity, caregiver education, household language, and food sufficiency. Survey-weighted logistic regression estimated class differences. Five classes emerged, with labels emphasizing the indicators that most differentiated profiles: Class 1 (52% of the population; high education, non-Hispanic [NH] White), Class 2 (10%; high education, NH Asian), Class 3 (5%; low education, NH White), Class 4 (20%; low education, Hispanic/Black), and Class 5 (12%; non-English, low education, Hispanic). The prevalence of meeting PA recommendations (19.48%) was lower than meeting SB recommendations (49.93%) and sleep recommendations (64.52%). Compared with Class 1, odds were lower in Classes 2 and 5 (ORs = 0.56–0.64) for meeting PA recommendations, lower in Classes 3–5 (ORs = 0.56–0.72) for meeting SB recommendations, and lower in all other classes (ORs = 0.49–0.76) for meeting sleep recommendations. Relative to Class 1, disadvantaged classes had higher odds of meeting none (ORs = 1.39–2.05) and lower odds of meeting all three (ORs = 0.48–0.63) recommendations. Findings highlight the need to consider combinations of individual characteristics and SDOH rather than one-size-fits-all approaches to improve 24HMB adherence across diverse groups.
Introduction
Daily movement behaviors—physical activity (PA), screen-based sedentary behavior (SB), and sleep—have emerged as key, modifiable lifestyle factors that collectively shape pediatric health (Tremblay et al., 2024). Robust evidence demonstrates that children and adolescents who engage in adequate PA, spend minimal time in recreational screen use, and obtain sufficient sleep have more favorable weight trajectories, cardiometabolic risk factors, mental health symptoms, and cognitive performance (Rollo et al., 2020; Tapia-Serrano et al., 2022). However, few U.S. children and adolescents meet recommendations for optimal health benefits in these domains (Carlson et al., 2025). According to the 2024 U.S. Report Card on PA for Children and Youth, only 20%–28% of children and adolescents ages 6–17 years meet recommendations to engage in at least 60 min of daily PA, and approximately 20% meet sedentary-related recommendations to limit recreational screen time to 2 hr or less per day (Carlson et al., 2025). Sleep patterns are somewhat more favorable. On weeknights, 64% of children and 67% of adolescents meet age-based minimum sleep duration recommendations, defined as 9–12 hr for ages 6–12 years and 8–10 hr for ages 13–18 years (Carlson et al., 2025). Furthermore, evidence reveals notable disparities in 24-hr movement behaviors (24HMB) across various socioeconomic and demographic groups (Carlson et al., 2025; Musić Milanović et al., 2021), indicating that children who are socially or economically disadvantaged may face additional barriers to achieving recommended movement behaviors. A child’s sex may also shape 24HMB patterns in behavior-specific ways. For instance, a recent analysis of national U.S. data found that female children and adolescents were less likely than their male counterparts to meet PA and all three 24HMB recommendations, yet were more likely to meet screen-time recommendations (Pfledderer et al., 2025).
There is increasing recognition within public health that an intersectionality lens, capturing how multiple social identities and positions interact, should be actively incorporated to inform population health promotion (Bauer et al., 2021). At the same time, a growing body of research indicates that social drivers of health (SDOH)—also known as social determinants of health, encompassing the conditions in which individuals are born, grow, live, work, and age—are key upstream contributors to pediatric health risk (Morris et al., 2024). In pediatrics, SDOH reflect family and broader socioeconomic contexts that shape child well-being (Morris et al., 2024). For example, exposure to adverse SDOH during childhood and adolescence may increase the risk of later cardiovascular and other chronic diseases (Viner et al., 2012). However, most prior 24HMB studies have examined disparities using single-factor comparisons and multivariable regression models (Bauer et al., 2021). Although these approaches are useful for estimating associations between specific characteristics and movement behaviors, they typically focus on variable-specific effects or require investigators to pre-specify interaction terms. Because intersectionality emphasizes that social identities and SDOH shape experience jointly rather than independently (Bauer et al., 2021), a person-centered approach can complement regression-based evidence by identifying latent profiles of children who share similar combinations of social identity and family-level SDOH indicators (Weller et al., 2020).
To address this gap, this study applied an exploratory latent class analysis (LCA) approach to identify patterns within a national sample of U.S. children and adolescents defined by the intersections of observed sociodemographic characteristics (race/ethnicity, sex) and family-level SDOH indicators (caregiver education, food sufficiency, and household language) (Bauer et al., 2021). Together, these indicators capture demographic identity, household socioeconomic resources, material hardship, and linguistic context that are recognized in health equity frameworks and may shape children’s opportunities to meet 24HMB recommendations (Morris et al., 2024; Teshale et al., 2023). LCA is a person-centered, data-driven technique that uncovers subpopulations with similar profiles across multiple indicators, offering a pragmatic way to apply an intersectionality framework (Weller et al., 2020). Compared with traditional approaches that focus primarily on independent covariate effects or selected interaction terms, LCA models the joint distribution of multiple indicators and can identify socially patterned subgroups that may be obscured in population-average estimates (Else-Quest & Hyde, 2016). This approach adds value to traditional regression-based approaches for informing interventions and policies by helping to show how combinations of social identity and family-level SDOH indicators are related to groups with distinct 24HMB needs.
The present study aimed to (a) identify latent subgroups of U.S. children and adolescents characterized by intersectional identities and (b) examine disparities in meeting 24HMB recommendations across subgroups. By explicitly accounting for how overlapping social contexts may relate to movement behaviors, this approach enhances our understanding of pediatric health disparities within actionable social contexts where health promotion efforts can be more effectively focused.
Methods
Data Source and Sample
This study used data from the 2022 to 2023 National Survey of Children’s Health (NSCH), an annual household survey providing national- and state-level information on the health and health care needs of children aged 0–17 (Child and Adolescent Health Measurement Initiative [CAHMI], 2025). We restricted analyses to children and adolescents aged 6–17 years. A parent or guardian completed the survey online or on paper, with telephone assistance as needed. Full documentation on the NSCH’s sampling design, data collection methodology, and measured variables is available elsewhere (Data Resource Center for Child and Adolescent Health, n.d.). The study adhered to the STROBE reporting guidelines for cross-sectional studies (von Elm et al., 2007). Because this study involved secondary analysis of publicly available, de-identified data, institutional review board review was not required.
Measures
Sociodemographic Indicators Used to Capture Intersectional Identities
This study focused on social identities related to cardiovascular disease burden and PA (Morris et al., 2024), specifically: the child’s biological sex (male or female) and race and ethnicity (Hispanic, Non-Hispanic [NH] White, NH Black, NH Asian, or other race/ethnicity [American Indian or Alaska Native, Native Hawaiian and Other Pacific Islander, or multiracial]), as well as SDOH indicators: caregivers’ educational level (less than high school, high school/some college, or college degree/higher), family food insufficiency (definitely sufficient, not always sufficient, sometimes/often insufficient), and primary household language (English or non-English) (Office of Disease Prevention and Health Promotion, n.d.; Morris et al., 2024). Primary household language was included as a proxy for acculturation (Morris et al., 2024). For ease of interpretation, race/ethnicity, educational level, and food insufficiency were recoded into the categories in parentheses.
24-hr Movement Behaviors
We identified 24HMB (PA, SB, and sleep) from survey items. Overall PA was assessed with, “During the past week, on how many days did this child exercise, play a sport, or participate in PA for at least 60 min?” (0 days, 1–3 days, 4–6 days, every day) and analyzed as a binary variable (meeting the 60 min every day recommendation vs. not) (U.S. Department of Health and Human Services, 2018; World Health Organization, 2020). SB (recreational screen time) was measured with, “On most weekdays, about how much time does this child usually spend in front of a TV, computer, cellphone, or other electronic device watching programs, playing games, accessing the internet, or using social media, not including schoolwork?” (<1 hr, 1 hr, 2 hr, 3 hr, ≥4 hr). This measure was recoded as a binary variable (meeting the ≤2 hr/day recommendation vs. not) (Tremblay et al., 2016). Sleep duration was assessed with, “During the past week, how many hr of sleep did this child get on most weeknights?” and coded using the NSCH-derived binary indicator for meeting age-specific minimum sleep duration recommendations (9–12 hr for ages 6–12 and 8–10 hr for ages 13–18) (Paruthi et al., 2016). Using the three binary indicators (PA, SB, sleep), we created an eight-level recommendation-combination variable (meeting none, PA only, SB only, sleep only, PA + SB, PA + sleep, SB + sleep, all three). We also created a four-level variable for the total number of recommendations met (none, exactly one of three, exactly two of three, all three), with participants categorized into mutually exclusive groups.
Covariates
Covariates related to children’s movement behaviors and health outcomes (Carlson et al., 2025; Fabricant et al., 2019; Fairclough et al., 2023; Marques et al., 2023; Morris et al., 2024) included age, body mass index (BMI)-for-age (underweight, normal weight, overweight or obese), and current health insurance status (insured during the past 12 months or at the time of the survey; yes/no).
Statistical Analysis
LCA of five sociodemographic factors was conducted in Mplus (version 8.10) using full-information maximum likelihood. Model selection was guided by a combination of statistical fit indices, including Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), adjusted BIC, entropy, class proportions, and theoretical interpretability to determine the optimal number and composition of classes (Weller et al., 2020). Individuals were assigned to the class for which they had the highest model-estimated probability of membership. This modal class assignment approach was used to support interpretable class-specific prevalence estimates and survey-weighted regression models while incorporating the NSCH complex survey design and multiple imputation procedures. Because modal assignment treats probabilistic class membership as a fixed class label, class-based estimates were interpreted as approximate group-level comparisons rather than definitive individual classifications.
All subsequent analyses were conducted in R (version 4.5.1) and accounted for the NSCH complex design via appropriate weighting, stratification, and clustering variables. Descriptive statistics were calculated to summarize sample characteristics and 24HMB recommendation adherence. We estimated weighted class-specific prevalence of meeting PA, screen-based SB, and sleep duration recommendations. Logistic regression models estimated adjusted odds ratios (ORs) for meeting recommendations for each individual behavior component and for meeting none, one, two, or all three 24HMB recommendations. ORs < 1 indicate lower odds than the reference class, defined as the most advantaged class. Statistical significance was assessed using Bonferroni-adjusted ≤.05 and 95% confidence interval (CI).
Item-level non-response was less than 5% for all analytic variables, within conventional thresholds for complete-case analysis and multiple imputation (Sinha et al., 2021). To minimize bias and maintain statistical power under a missing-at-random assumption (Pedersen et al., 2017; Rubin, 1996; Sinha et al., 2021), we used multiple imputation by chained equations (MICE; m = 10 imputed data sets). After imputation, we constructed 24HMB combinations in each imputed data set, fit models separately, and combined estimates using Rubin’s rules (van Buuren & Groothuis-Oudshoorn, 2011). Because the public-use NSCH file hot-deck imputes weighting variables (child sex, race/ethnicity) and sequentially imputes adult education, our imputation strategy built on this framework to address the remaining analytic variables. As a sensitivity analysis, we compared imputed and complete-case results (Sinha et al., 2021).
Results
Sample Characteristics
Table 1 summarizes sample characteristics (unweighted N = 68,000; weighted mean age 11.61 years, SD 3.43). Overall, 51.23% were male, and NH White was the largest racial/ethnic group (47.28%). Among caregivers, 9.27% had less than a high school education, 39.21% had completed high school/some college, and 51.52% had a college degree or higher. Most households were English-speaking (84.60%), and 66.06% reported food sufficiency. Health insurance coverage was 93.10%. Overweight/obesity prevalence was 32.16%.
Sample Characteristics of U.S. Children and Adolescents.
Note. Descriptive statices were done using complete case. H = non-Hispanic.
Weighted mean and standard deviation.
Latent Intersectional Subgroups
A five-class model was selected because it had the lowest BIC and adjusted BIC (BIC = 461,471; ABIC = 461,300). The six-class model did not improve these criteria and had lower entropy than the five-class model (0.69 vs. 0.77; Supplementary Table S1). The Vuong–Lo–Mendell–Rubin and bootstrapped likelihood ratio tests supported the five-class model over the four-class model (p < .001). Together, these results supported the five-class solution as having favorable model fit and reasonable classification quality. Figure 1 depicts the unweighted latent class profiles in the observed sample. Based on modal class assignment, Class 1 represented the largest unweighted proportion of the analytic sample (65.2%), followed by Class 3 (11.7%), Class 4 (11.2%), Class 2 (6.3%), and Class 5 (5.6%). The corresponding survey-weighted class proportions are presented in Table 2. Class 1 remained the largest class in the weighted distribution, reflecting 52% of the population.
Heatmap of five-class latent class analysis.
Distribution of 24-hr Movement Behaviors Across Latent Classes.
Note. Estimates are pooled across multiply imputed data sets; values are presented as unweighted counts (N) and weighted percentages. PA = physical activity; SB = sedentary behavior; Class 1 (reference) = high-education, food-sufficient, NH White; Class 2 = high-education, food-sufficient, NH Asian; Class 3 = low-education, non-Hispanic (NH) White; Class 4 = English, low-education, NH Hispanic or Black; Class 5 = non-English language, low-education, Hispanic.
Class 1 emerged as the more socially advantaged group (80.2% NH White children and adolescents; 76.3% of caregivers completed ≥ college; 99.7% English-speaking households; 89.1% rated food sufficiency as definitely). Class 2 was also advantaged (63.0% NH Asian children and adolescents; 87.5% of caregivers completed ≥ college; 62.1% English-speaking households; 89.0% rated food sufficiency as definitely). Classes 3–5 were relatively disadvantaged. Class 3 mostly comprised NH White children and adolescents (75.8%), caregivers without a college degree (76.4%), English-speaking households (100.0%), and families with lower food sufficiency (83.5% not always; 16.5% sometimes/often insufficient). Class 4 included Hispanic children and adolescents (52.1%) and NH Black children and adolescents (34.4%), caregivers without a college degree (73.4%), English-speaking households (95.8%), and families with moderate-to-low food sufficiency (46.7% definitely; 41.0% not always; 12.3% sometimes/often insufficient). Class 5 predominantly comprised Hispanic children and adolescents (80.2%), caregivers without a college degree (84.5%), non-English-speaking households (100.0%), and moderate-to-low food sufficiency (53.6% definitely; 38.7% not always; 7.8% sometimes/often insufficient). Sex was included as an LCA indicator but did not strongly distinguish the five classes. Across classes, the probability of being male ranged from 50.8% to 52.9%, and the probability of being female ranged from 47.1% to 49.2%, indicating that the profiles were primarily differentiated by race/ethnicity, caregiver education, food sufficiency, and household language (Supplementary Table S2).
Prevalence of Meeting 24-hr Movement Behavior Recommendations
Table 2 presents the prevalence of meeting the 24HMB recommendations among each latent class. The prevalence of meeting PA recommendations was low across all classes (14–22%), highest in Class 3 (21.6%) and Class 1 (21.1%) and lowest in Class 5 (14.0%) and Class 2 (13.6%). The prevalence of meeting SB recommendations was highest in Class 1 (55.6%) and Class 2 (55.4%) and lowest in Class 4 (40.2%). The prevalence of meeting sleep recommendations was highest in Class 1 (71.5%) and lowest in Class 3 (53.7%) and Class 4 (53.6%). The prevalence of meeting all three recommendations was largest in Class 1 (11.8%) and only 6–7% in the other classes, whereas meeting none was most common in Classes 4 (24.9%) and 3 (23.5%) and least common in Class 1 (13.0%).
Class Differences in Meeting 24HMB Recommendations
Table 3 demonstrates distinct patterns in the odds of meeting recommendations for each 24HMB across latent classes, using Class 1 as the reference. For PA, Class 2 (OR = 0.56, 95% CI = 0.48–0.65) and Class 5 (OR = 0.64, 95% CI = 0.54–0.76) had significantly lower odds of meeting PA recommendations (both p < .001), whereas the odds for Class 3 (OR = 1.09, 95% CI = 0.98–1.23, p = .13) and Class 4 (OR = 0.97, 95% CI = 0.86–1.09, p = .57) did not differ from Class 1. For SB, Classes 3 (OR = 0.59, 95% CI = 0.54–0.65), 4 (OR = 0.56, 95% CI = 0.51–0.62), and 5 (OR = 0.72, 95% CI = 0.63–0.82) had lower odds of meeting SB recommendations than Class 1 (all p < .001), while the odds for Class 2 and Class 1 were not significantly different (OR = 0.96, 95% CI = 0.85–1.07, p = .46). For sleep, all comparison classes had lower odds of meeting sleep recommendations than Class 1 (all p < .001), with the largest differences for Class 3 (OR = 0.49, 95% CI = 0.44–0.53) and Class 4 (OR = 0.49, 95% CI = 0.45–0.53), followed by Class 5 (OR = 0.67, 95% CI = 0.59–0.76) and Class 2 (OR = 0.76, 95% CI = 0.67–0.85).
Associations Between Latent SDOH Classes and Adherence to Each Individual 24-hr Movement-Behavior Guideline.
Note. Bonferroni-corrected significance threshold = .004. SE = standard error; OR = odds ratio; CI = confidence interval; PA = physical activity; SB = sedentary behavior; Class 1 (reference) = high-education, food-sufficient, NH White; Class 2 = high-education, food-sufficient, NH Asian; Class 3 = low-education, non-Hispanic (NH) White; Class 4 = English, low-education, NH Hispanic or Black; Class 5 = non-English language, low-education, Hispanic.
Differences across classes were also observed in the number of 24HMB recommendations met (Table 4). As compared to those in Class 1, Classes 2–5 had higher odds of meeting none of the three recommendations (ORs = 1.39–2.05; all p < .001) and lower odds of meeting all three recommendations (ORs = 0.48–0.63; all p < .001). For meeting exactly one recommendation, Classes 3–5 had higher odds than Class 1 (ORs = 1.20–1.29, all p < .001), whereas Class 2 did not differ from Class 1 (OR = 1.07, 95% CI = 0.95–1.21, p = .25). Classes 3–5 had lower odds of meeting exactly two recommendations (ORs = 0.62–0.75; all p < .001), whereas Class 2 did not differ from Class 1 (OR = 0.97, 95% CI = 0.86–1.09, p = .60).
Associations Between Latent Classes and 24-hr Movement Behavior Combinations.
Note. Bonferroni-corrected significance threshold = .003. Class 1 (reference) = high-education, food-sufficient, NH White; Class 2 = high-education, food-sufficient, NH Asian; Class 3 = low-education, non-Hispanic (NH) White; Class 4 = English, low-education, NH Hispanic or Black; Class 5 = non-English language, low-education, Hispanic; Overall adherence categories are mutually exclusive.
Discussion
To the best of our knowledge, this is one of the first studies to use a national sample of U.S. children and adolescents to explore intersectional identity subgroups and examine disparities in meeting 24HMB recommendations. Five distinct latent classes captured overlapping advantages and disadvantages. The two relatively advantaged classes were Class 1 (high-education, food-sufficient, NH White) and Class 2 (high-education, food-sufficient, NH Asian), while children and adolescents in the three relatively disadvantaged classes were more likely to exhibit intersecting factors related to caregiver education (Classes 3, 4, and 5 having low caregiver education), race/ethnicity (Class 4 being largely Hispanic or NH Black and Class 5 being largely Hispanic), and food sufficiency, with all three having higher rates of food insufficiency. The prevalence of meeting 24HMB recommendations differed across classes. At least one relatively advantaged class had a higher prevalence, while relatively disadvantaged classes generally had lower prevalence, although the rank order differed for meeting the PA, screen-based SB, and sleep recommendations. These behavior-specific differences suggest that class profiles may have distinct 24HMB needs. Notably, the emergence of two Hispanic-relevant profiles with different social characteristics and patterns of meeting recommendations suggests that important heterogeneity may be less visible when using broad racial/ethnic categories alone. The findings suggest that 24HMB disparities are shaped by combinations of social identity and family-level SDOH rather than single characteristics alone, supporting targeted strategies responsive to the social and household contexts reflected in each profile alongside population-wide strategies.
The comparison between Classes 4 and 5 illustrates this heterogeneity. Hispanic children and adolescents were not represented by a single homogeneous profile, and Classes 4 and 5 revealed differences that would not be observable when focusing on race/ethnicity or caregiver education individually. Class 4 included a large Hispanic subgroup together with NH Black children and adolescents and was largely English-speaking with lower caregiver education, whereas Class 5 was predominantly Hispanic, non-English speaking, and had the lowest caregiver education profile. Interestingly, they showed different 24HMB patterns, with only Class 5 experiencing disparities in meeting PA recommendations (compared to Class 1). These findings suggest that subgroups within minoritized racial/ethnic groups may experience larger health disparities, particularly among Hispanic children and adolescents. Thus, when only racial/ethnic characteristics are considered, we are likely to underestimate disparities experienced by subgroups with intersecting SDOH, particularly among Hispanic children and adolescents in families with low caregiver education and a non-English primary household language, which may reflect different language- and acculturation-related contexts. Of note, the prevalence of meeting PA recommendations was slightly higher in Class 4 compared with prior estimates for NH Black children and adolescents (Carlson et al., 2025; Nagata et al., 2022), suggesting that the higher prevalence in Class 4 may be more reflective of the Hispanic subgroup than the NH Black subgroup. Overall, these findings from the two largely Hispanic latent classes suggest that health promotion strategies should be designed around intersections of social identities and SDOH, with increased focus on identifying subgroups with the highest risk profile.
The rank order from highest to lowest prevalence was similar for meeting the SB and sleep recommendations. Prevalence was highest in the two advantaged classes (Class 1: NH White, high education and Class 2: NH Asian, high education), in the middle for Class 5 (Non-English speaking, Hispanic, low-education), and lowest in the other disadvantaged classes (Class 3: White, low-education; Class 4: Hispanic/Black, low-education), with similar magnitudes of differences between the highest and lowest classes (~15–18 percentage points). These population-level gaps highlight the need for targeted improvement efforts. However, the pattern for meeting the PA recommendation was distinct. Class 2, which had relatively high prevalence of meeting the SB and sleep recommendations, had the lowest prevalence of meeting the PA recommendation, with Class 5 having similarly low prevalence of meeting the PA recommendation. Class 3, which had low prevalence of meeting the SB and sleep recommendations, had the highest prevalence of meeting the PA recommendation, with Class 1 having similarly high prevalence of meeting PA recommendations and Class 4 having a prevalence just below Classes 1 and 3. These patterns should be interpreted with two class-profile considerations in mind. First, although sex was included as an LCA indicator, it did not strongly differentiate the latent profiles. This suggests that the selected classes were primarily shaped by race/ethnicity and family-level SDOH indicators rather than by sex, although sex and gender-related norms may still shape daily expectations, opportunities, and routines related to PA, screen-based SB, and sleep within or across class profiles (Lee et al., 2023). Second, because Class 1 accounted for the largest share of the weighted class distribution, overall weighted prevalence estimates may partly reflect the movement-behavior patterns of this relatively advantaged class, reinforcing the value of class-specific comparisons.
The relatively favorable pattern of meeting the PA, SB, and sleep recommendations in Class 1 aligns with prior evidence showing differences by race/ethnicity and caregiver education level (Carlson et al., 2025). However, even in this advantaged profile, 78.9% did not meet PA recommendations, indicating that physical inactivity remains a public health concern and that population PA levels have not improved over the past decade (Carlson et al., 2025). The lower prevalence of meeting PA recommendations in Class 2 is consistent with prior studies showing 14% of Asian children and adolescents meeting PA recommendations (Carlson et al., 2025). One possible but untested explanation is that time-use or academic-context factors may reduce opportunities for recreational or structured PA in this high caregiver education, NH Asian profile (Dunatchik & Park, 2022; Kim, 2021). However, these factors may plausibly affect children and adolescents across multiple racial/ethnic and household contexts, including households with high caregiver education. If future studies confirm time-related or academic-context barriers, brief PA opportunities integrated throughout the day (e.g., classroom activity breaks, active learning, recess activities, intramurals) may be useful for children and adolescents facing these constraints, rather than programs requiring longer uninterrupted periods that may be perceived as burdensome or as competing with academic goals. Given heterogeneity within Asian populations (e.g., caregiver education differs between Asian Indian and other subgroups) (Min et al., 2022), health promotion efforts should consider cultural and subgroup differences.
For Classes 3 and 4, the prevalence of meeting the screen-based SB and sleep duration recommendations was consistently lower than in the relatively advantaged groups. These patterns align with the class-defining indicators of lower caregiver education and lower food sufficiency, which may reflect fewer household resources to support consistent screen-time limits, sleep routines, and low-cost, safe alternatives to recreational screen use (Rhodes et al., 2020; Tandon et al., 2012). These findings underscore the need for interventions and policies that include accessible caregiver-facing guidance, practical sleep-hygiene strategies, reduced pre-bedtime screen exposure, and low- or no-cost opportunities for structured PA activities. Broader contextual factors (e.g., caregiver work schedules, housing conditions, neighborhood safety, and community program availability) may also shape SB and sleep patterns, but they were not directly measured in this study and should be considered future research priorities or intervention settings.
Class 5 showed generally low prevalence of meeting recommendations across all three behaviors, although the prevalence of meeting the SB and sleep recommendations was not the lowest. This predominantly Hispanic class was characterized by a non-English primary household language and the lowest caregiver education. Because household language is linked with acculturation (Morris et al., 2024; Sentell & Braun, 2012), this pattern may indicate a lower language- and acculturation-related context than Hispanic households in Class 4. Acculturation has been associated with higher PA among Hispanic/Latino children and adolescents, and parents with greater Anglo‑oriented acculturation are more likely to use positive parenting practices (e.g., monitoring/reinforcement) linked to higher MVPA (Gonzalez et al., 2023). As in Classes 3 and 4, lower caregiver education, combined with non-English language preference or limited English proficiency, may increase PA barriers partly through lower health literacy and reduced ability to navigate PA opportunities (e.g., organized sports and community resources) (Santana et al., 2021). First-generation immigrant households may face added challenges accessing health information in a non-native language, and among Latinos, limited English proficiency (LEP) or low health literacy has been associated with poorer self-rated health (Sentell & Braun, 2012). However, our results should be interpreted cautiously, as primary household language may not fully capture English proficiency, and maintaining the home language can support family cohesion, ethnic identity, social support, and socioemotional well-being.
Beyond language and health literacy, structural and contextual factors may contribute to low prevalence of meeting PA recommendations in Class 5 (Larsen et al., 2013). Predominantly Latino neighborhoods with greater socioeconomic disadvantage may have fewer safe, affordable places for children and adolescents to be active and fewer school‑ or community‑based PA programs. Families may also face greater transportation and cost barriers, immigration‑related stressors, complex institutional requirements, and long or irregular work schedules that limit caregiver support for PA. Culturally and linguistically tailored, asset-based programs that build on family involvement, collectivism, and intergenerational support, alongside free or low-cost opportunities and stronger neighborhood PA resources in lower-income communities that have faced decades of disinvestment, may improve reach and effectiveness for this profile.
Meeting all three 24HMB recommendations was low—only 9.19% of children and adolescents met all three recommendations, and 17.51% met none of the three recommendations. This aligns with previous national estimates (Pfledderer et al., 2025), indicating persistent gaps in meeting 24HMB recommendations. Notably, meeting none of the recommendations was more common in the disadvantaged classes (20.0–24.9% vs. 16.8%), and meeting all three recommendations was rare overall but higher in the advantaged class (11.8% vs. 6–7%). This clustering of unmet recommendations suggests potential compounding risk and supports multicomponent interventions, especially because optimizing all three behaviors is associated with better health outcomes and synergistic effects (Kracht et al., 2024). Such efforts may be particularly important among children and adolescents in households with low caregiver education, which often co-occurs with lower income.
This study has several notable strengths. First, it advances 24HMB research by applying an intersectionality-informed, person-centered framework to uncover subgroups defined by overlapping social identities and SDOH, providing a richer understanding of 24HMB disparities than single-variable analyses. For instance, within NH White households, movement patterns varied by caregiver education, and among Hispanic households, those with a non-English primary household language consistently had low movement profiles, whereas English-speaking households showed different PA patterns. This approach helps lay the groundwork for precision PA and health promotion strategies tailored to intersecting groups. Second, we leveraged the most recent, nationally representative NSCH waves and accounted for the complex survey design, which minimized bias in prevalence estimates and standard errors and enhanced generalizability. Finally, the large sample size supported stable class estimation and the detection of meaningful disparities across the five latent classes (Finch & Bronk, 2011).
While this study offers important insights into the intersection of social identities, SDOH, and movement behaviors among U.S. children and adolescents, several limitations should be noted. First, the cross-sectional design limits causal inferences between latent intersectional profiles and meeting 24HMB recommendations. Second, the measures used in this study were caregiver-reported and may be subject to recall bias. Third, our LCA was based on a selected set of social identity and family-level SDOH indicators and did not include potentially relevant constructs such as neighborhood safety, housing security, built environment, or household income level, suggesting that additional heterogeneity may remain within the population, including within the large Class 1 profile. However, many SDOH indicators tend to be interrelated, and the included variables in our study (e.g., education, food sufficiency) were able to capture key constructs that shape children’s daily movement opportunities. Relatedly, structural and contextual factors discussed as possible explanations should be interpreted as plausible contextual pathways. Finally, LCA is an exploratory, person-centered method, and class assignment is inherently probabilistic rather than fixed (Lanza & Rhoades, 2013; Weller et al., 2020). We used modal class assignment, which assigns each individual to the class with the highest model-estimated probability of membership, to facilitate interpretable class-specific prevalence estimates and survey-weighted regression models. However, this approach does not fully account for uncertainty in class membership and may make between-class differences appear more distinct or influence estimated class-outcome associations. While our quantitative intersectionality approach to examining 24HMB disparities goes beyond single-axis analyses, the findings should be interpreted as model-based approximations rather than definitive representations of individuals’ lived experiences of intersecting social contexts (Weller et al., 2020). Future studies should incorporate objective, accelerometer-based measures and meaningful contextual data (e.g., neighborhood, built environment, school) to strengthen inference and reduce measurement errors. Investigators should also examine mechanisms underlying class-specific behavior patterns within a socioecological framework to elucidate environmental and psychosocial influences. Such work should also consider how sex and gender-related expectations, norms, and opportunities interact with race/ethnicity and family-level SDOH to shape daily movement routines. Factors that moderate or protect against the detrimental impacts that negative SDOH can have on health behaviors should also be further investigated to support efforts to mitigate these impacts. Longitudinal analyses are needed to evaluate how intersectional characteristics/contexts shape 24HMB and health trajectories over time, and to test whether movement behaviors mediate or moderate these associations. In addition, surveys that oversample underserved and marginalized populations are needed to provide deeper insight into contemporary barriers and facilitators and to ensure adequate statistical power for subgroup analyses.
Conclusion
The patterns of 24HMB disparities across the latent intersectional profiles identified in this national sample suggest that many children and adolescents face overlapping structural, economic, and social barriers that limit access to resources and opportunities. Since meeting all three recommendations was low across all classes, population-wide efforts to improve 24HMB (particularly PA) remain necessary. Policy-level levers may include equity-focused school wellness policies that protect access to high-quality physical education, recess, classroom activity breaks, and after-school programs, particularly in schools serving socially or economically disadvantaged students, as well as local planning policies that improve the safety, accessibility, and quality of parks, sidewalks, and active routes to school in lower-resource communities. These population-wide approaches should be paired with targeted supports responsive to the social and household contexts reflected in U.S. population groups, including accessible caregiver-facing guidance for families with lower caregiver education, low- or no-cost PA and non-screen activity opportunities for families experiencing material constraints, linguistically responsive communication for households with a non-English primary language, and flexible PA opportunities that can be integrated into school and daily routines for groups with low prevalence of meeting PA recommendations. The emergence of distinct Hispanic-relevant profiles further suggests that household language, caregiver education, and material resource context should be considered when tailoring supports within broad racial/ethnic groups. In this study, profiles showing lower prevalence across multiple 24HMB recommendations, particularly the predominantly Hispanic profile characterized by non-English primary household language and lower caregiver education, may warrant early attention in local planning, while the lower-caregiver education NH White and Hispanic/NH Black profiles highlight the need for domain-specific supports for screen-based SB and sleep duration. Because intervention priorities will vary by local population composition and available resources, practitioners should use local needs assessments to identify which intersectional profiles and 24HMB domains require the most immediate support. Although infrastructure and access-related factors were not directly measured in this study, sustainable funding to implement evidence-based 24HMB strategies, school- and community-based infrastructure, and partnerships that expand safe, affordable opportunities for movement and sleep-supportive routines may help initiate and maintain these efforts in communities experiencing resource constraints.
Supplemental Material
sj-docx-1-heb-10.1177_10901981261466988 – Supplemental material for Intersectional Identities and 24-hr Movement Behavior Recommendation Adherence Among U.S. Children and Adolescents
Supplemental material, sj-docx-1-heb-10.1177_10901981261466988 for Intersectional Identities and 24-hr Movement Behavior Recommendation Adherence Among U.S. Children and Adolescents by Suryeon Ryu, Jordan A. Carlson and Helena H. Laroche in Health Education & Behavior
Footnotes
Author’s Note
Jordan A. Carlson and Helena H. Laroche are also affiliated with the University of Missouri–Kansas City, Kansas City, Missouri, USA.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
