Associations of Changes in Sleep and Emotional and Behavioral Problems From Late Childhood to Early Adolescence

Abstract

Importance: Sleep problems and psychopathology symptoms are highly comorbid and bidirectionally correlated across childhood and adolescence. Whether these associations are specific to discrete profiles of sleep problems and specific internalizing and externalizing phenomena is currently unclear.

Objective: To characterize individual changes in profiles of sleep problems and their prospective associations with psychopathology symptoms across the transition from childhood to adolescence.

Design, setting, and participants: This observational cohort study used baseline data (participant age of 9 to 11 years) and 2-year follow-up data (participant age of 11 to 13 years) from the community-setting, multicenter Adolescent Brain Cognitive Development (ABCD) study. Individuals were assessed for a range of sleep problems at both waves and categorized into profiles via latent profile analysis. The stability and change in these profiles over time was assessed via latent transition analysis. Logistic regression models examined whether psychopathology symptoms were cross-sectionally associated with profile membership and whether transitions between profiles were associated with changes psychopathology symptoms over time. Data were collected from September 2016 to January 2020, and data were analyzed from August 2021 to July 2022.

Exposures: Sleep problems were assessed at both baseline and follow-up via the parent-reported Sleep Disturbance Scale for Children (SDSC).

Main outcomes and measures: Psychopathology symptoms at both baseline and follow-up were assessed using the internalizing and externalizing dimension scores derived from the parent-reported Child Behavior Checklist.

Results: A total of 10 313 individuals (4913 [47.6%] were female) were categorized into 4 latent profiles of sleep problems at both baseline and follow-up: a low disturbance profile, a sleep onset/maintenance problems profile, a moderate and nonspecific disturbance profile (termed mixed disturbance), and a high disturbance profile. Individuals in the 3 more severe problem profiles displayed greater risk of concurrent internalizing symptoms (sleep onset/maintenance problems: odds ratio [OR], 1.30; 95% CI, 1.25-1.35; P < .001; mixed disturbance: OR, 1.29; 95% CI, 1.25-1.33; P < .001; high disturbance: OR, 1.44; 95% CI, 1.40-1.49; P < .001) and externalizing symptoms (sleep onset/maintenance problems: OR, 1.20; 95% CI, 1.16-1.23; P < .001; mixed disturbance: OR, 1.17; 95% CI, 1.14-1.20; P < .001; high disturbance: OR, 1.24; 95% CI, 1.21-1.28; P < .001). Transitions between sleep profiles over time were associated with prospective internalizing and externalizing symptoms, but not vice versa.

Conclusions and relevance: There are substantial changes in sleep problems across the transition to adolescence that are associated with later internalizing and externalizing symptoms. Sleep profiles could be targeted in future intervention and treatment programs to improve sleep-related and mental health-related outcomes across development.

Figure 1.. Sleep Problem Profiles in Late Childhood and Their Change Over Time

A, Mean scores on subscales of the Sleep Disturbance Scale for Children (SDS) for each sleep profile in late childhood. All subscales were converted to z scores to enable comparisons across scales. Note that duration is reverse scored, such that greater scores indicate shorter sleep duration. B, Transitions between sleep profiles from late childhood (top) to young adolescence (bottom). Percentages and line thickness represent the proportions of each profile that remained in the same class over time or transitioned to a different profile. Note that no individuals transitioned from the sleep onset/maintenance problems profile into the mixed disturbance profile, from the low disturbance profile into the high disturbance profile, or the high disturbance profile into the low disturbance profile over time (ie, proportions were 0%).

Figure 2.. Cross-sectional Associations Between Sleep Profiles and Psychopathology Symptoms

A, Profile membership in relation to internalizing and externalizing scores. B, Profile membership in relation to specific internalizing and externalizing symptoms. Odds ratios (ORs) represent the association between each Child Behavior Checklist (CBCL) dimension score and the probability of profile membership compared with the low disturbance profile. Models were adjusted for age, sex, race and ethnicity, socioeconomic status, and body mass index. Error bars indicate 95% CIs.

Figure 3.. Longitudinal Associations Between Change in Sleep Profiles and Changes in Internalizing and Externalizing Symptoms

Longitudinal change in dimensional Child Behavior Checklist (CBCL) scores over time for each sleep disturbance transition pattern. Positive scores indicate increases in symptoms, and negative scores indicate decreases in symptoms. Error bars indicate 95% CIs. Footnotes indicate whether changes in scores are significantly different from zero, as tested with 1-sample t tests. ^aP < .001. ^bP < .05.

Figure 4.. Longitudinal Associations Between Change in Sleep Profiles and Changes in Child Behavior Checklist (CBCL) Subscale Scores

Longitudinal change in subscale CBCL scores over time for each sleep disturbance transition pattern. Positive scores indicate increases in symptoms, and negative scores indicate decreases in symptoms. Error bars indicate 95% CIs.