Background: Excessive screen time (>2 hours/day) is associated with childhood overweight and obesity and unfavorable obesogenic behaviors such as physical inactivity, sedentary time, and disrupted sleep. Prior reviews indicate intervening on screen time leads to improvements in screen time and other obesogenic behaviors; yet it is unclear what behavioral strategies and intervention characteristics are most important to consider in the design of screen time interventions.

Purpose: A systematic review and meta-analysis was conducted to identify behavioral strategies and intervention characteristics associated with treatment effectiveness in behavioral interventions targeting reductions in children’s screen time.

Methods: A computer-based search strategy employing keyword and controlled vocabulary terms with four databases (Web of Science, EMBASE, Ebscohost, Pubmed) was executed during January-February 2020. Behavioral interventions targeting reductions in children’s (0-18 years) screen time were eligible for inclusion. Intervention characteristics (e.g., sample size, duration) and the type of behavioral strategies used (e.g., education, social support, goal-setting, or accountability) were extracted.

Results: 11,949 articles were reviewed for inclusion based on title/abstract of which 595 underwent full-text review. Of these, 216 underwent narrative extraction and 186 were included in the random-effects meta-analysis. The overall standardized mean difference (SMD) was 0.12 (95%CI 0.08 to 0.16), suggesting small improvements in children’s screen time. Across all studies, incorporating goal strategies was associated with larger effects (SMD=0.18, 0.13 to 0.23) vs. studies without these strategies (SMD=0.04, -0.02 to 0.10). No other strategies were associated with intervention effectiveness. Accounting for sample size, goal-setting strategies were most effective in small studies (N<95, SMD=0.31, 0.22 to 0.41) with this effect dissipating with progressively larger studies (N>697, SMD=0.08, -0.02 to 0.17). In the absence of goal-setting strategies, effectiveness was associated with sample size (N<95 SMD=0.20, 0.01 to 0.40; N>697 SMD=0.06, -0.05 to 0.16).

Conclusion: Goal-setting strategies result in greater reductions in children's screen time. However, this impact is predominately driven by smaller studies. As sample size increased, studies incorporating goal strategies were no more effective than those studies that did not. Identifying ways to maintain the effectiveness of promising behavioral strategies identified in small trials when scaled are needed.

Background: Screen time (ST) is known to disrupt children’s sleep duration and timing, both acutely and chronically. However, few studies have examined the reverse effect of sleep on subsequent ST, particularly in elementary-aged children. Insufficient sleep may contribute to next-day fatigue and predispose children toward engaging in less energy-demanding activities, such as ST. Children’s sleep may be regulated by household routines, bedtime rules, and structured activities. On less-structured days (e.g., weekends), children may modify their behaviors, by sleeping longer, shifting their bed and wake times later, and engaging in sedentary leisure-time activities, such as increased ST. This study examined the temporal association between children’s sleep duration and next-day ST.

Methods: Parents of 200 children (6-12 years, 50% female, 36% white) reported their child’s daily bedtime, wake time, and total ST over 5 waves of 30-day data collection sprints between 2018-2019. Sleep duration was calculated as the difference between children’s wake time and previous bedtime. A multilevel linear mixed model (observations nested within children) was used to estimate total ST from the previous night’s sleep duration, accounting for time in the study, day of week, gender, age, and race. Sleep duration was mean-centered to estimate between- and within-child effects.

Results: On average, children had more total ST (B=20.1 minutes, 95%CI 12.9, 27.3) on weekends compared to weekdays. Within-subjects effects revealed that on nights when children slept an hour longer than usual, their total ST the following day was 6.6 minutes (95%CI 3.7, 9.5) higher than their usual total ST. Between-person effects showed that on average, children who slept the longest did not have significantly more total ST than their peers (B=-8.7, 95%CI -18.6, 1.2).

Conclusions: While previous research has linked ST to subsequent disrupted sleep, this study explored the reverse relationship and found children had more ST following nights of longer sleep, even after accounting for day of week. Future analyses should explore the link between increased sleep duration and next-day ST on weekdays versus weekend days during the school year and summer. Further, the regulatory effects of parental rules and routines for bedtime on children’s sleep and ST is warranted.

Background: Approximately 1.5 billion elementary-age children were impacted by school closures and began virtual learning due to the COVID-19 pandemic during March/April 2020. Cross-sectional and retrospective longitudinal studies indicate that children’s screen time increased during the pandemic. However, they do not account for expected maturation changes. This Interrupted time-series design compares changes in children’s screen time from 2019-2020 (Pandemic) to changes from 2018-2019.

Methods: Parents of 231 elementary school students (ages 7-12) from a larger cohort reported children’s total screen time, evening screen time (after 8:00pm), and bedroom screen time (after 8:00pm in the bedroom) on 2-3 random days each week for 6 weeks in the Spring (April/May) and Summer (June/July) from 2018-2020. Three multilevel mixed models were used to estimate differences in means and changes in slope between years, accounting for age, sex, and race. 

Results: Prior to the pandemic (2018-2019), all screen time categories increased during both spring [total (B=22.05, 95%CI (9.84, 34.27), evening (B=19.18, 95%CI (12.50, 25.85), bedroom (B=6.76,  95%CI (1.54, 11.98)] and summer [total (B=30.67, 95%CI (19.39, 41.95), evening (B=20.85, 95%CI (14.68, 27.02),  bedroom (B=8.60,  95%CI (3.77, 13.43)]. During the pandemic, a significant acceleration occurred from 2019-2020 compared to 2018-2019 for total screen time during the spring (B=97.21, 95%CI (79.00, 115.43), and summer (B=18.27, 95%CI (1.12, 35.41), accounting for 97.2 ± 9.3 and 18.3 ± 8.8 additional minutes of screen time, respectively. However, accelerated changes in evening and bedroom screen time were not detected during the pandemic compared to the rate of change from 2018-2019 during the spring [evening (B=9.74, 95%CI (-0.03, 19.52), bedroom (B=3.92, 95%CI (-3.71, 11.55)] or the summer [evening (B=-7.05, 95%CI (-16.37, 2.28), bedroom (B=3.90, 95%CI (-3.39, 11.19)]. 

Conclusion: Virtual learning likely contributed to the observed increase in total springtime screen use during the pandemic. However, continued elevated screen use in summertime is concerning. It appears that screen time patterns established in the spring persisted into the summer, even in the absence of virtual learning. This study did not differentiate between academic and non-academic screen use; future studies are needed to evaluate if academic screen time differentially impacts health outcomes.

Purpose: Poor sleep quality during childhood is associated with obesity and other chronic diseases. Further exploration of the association between screen time exposure (particularly before bedtime) and poor sleep is needed. In addition, parenting practices are an important determinant of children’s health, and psychosocial predictors such as parental self-efficacy impact parenting behavior and child health outcomes per recent research. Considering that Latinx children have the highest prevalence of obesity in the United States and that Latinx families may have increased exposure to screen time, this study aims to examine the association between screen time parenting self-efficacy, screen time exposure, sleep among Latinx children.

Methods: This was a cross-sectional study with 101 Latinx parents of 2-to-5-year old children. Parents completed a survey (in-person and online) that measured: (1) Screen time parenting self-efficacy (adapted from a validated questionnaire) (2) Screen time in the hour before bedtime across devices (e.g. television and mobile media devices – binary “Yes/No”), and (3) Parent-reported child sleep duration. Analysis approaches include logistic regression, ANOVA, and hierarchical linear regression controlling for covariates.

Results/Findings: Children averaged 46 months of age(SD = 13), 49% were girls, and 44% were overweight/obese. Our main findings were: (1) Higher screen time parenting self-efficacy was associated with lower odds of exposure to: (a) mobile media devices (Exp(B) = .635, p = .023), (b) video games (Exp(B) = .534, p = .009), (c) PC/laptops (Exp(B) = .590, p = .021) in the hour before bedtime. (2) Children exposed to (i) video games, (ii) PC/laptops, (iii) mobile media devices, and (iv) TV in the hour before bedtime had significantly lower average daily sleep duration (range of p from <.001 to = .044). (3) screen time parenting self-efficacy accounted for child sleep duration above and beyond the effects of screen time exposure across different devices (range of p from .006 to .017).

Conclusions: Lower screen time parenting self-efficacy and screen time exposures before bed were associated with lower sleep duration among our sample of Latinx children. Further research examining screen time parenting and exposure among Latinx children are needed.

Purpose: Excessive TV-viewing among children is a public health concern, yet tools to measure children’s TV viewing suffer from biases. Our goal was to develop FLASH-TV, an objective measure of children’s TV viewing using computer vision and machine learning algorithms to analyze video images of the child in front of a TV.

Methods: Four design studies were done with family triads (parent and 2 siblings). Three were in an observation lab and 1 in the child’s home. The FLASH-TV system included a video camera placed near the TV facing the room in front of the TV. As a gold standard, video data were coded by staff using duration coding at the frame level (~30 frames/sec) for whether the target child’s gaze was on the TV (10% double coded, mean Kappa 0.83-0.88). FLASH-TV estimates a child’s TV viewing time by first detecting faces in a frame, next verifying that the face is the target child, and last assessing TV-watching (gaze) behavior. Each step is based on modifications of convolutional neural network algorithms. The target child’s TV viewing time estimated by FLASH-TV running the three steps sequentially was compared to the gold standard. Criterion validity for overall TV viewing time and frame-by-frame gaze between FLASH-TV and gold standard was calculated using intra-class correlation (ICC) in a generalized linear mixed model framework.

Results: 21 parent-sibling triads participated in one of 4 design studies. The children’s mean age was 10.2 years, with 38.1% non-Hispanic white, 28.6% black, 19% Hispanic white, and 14.4% other. The ICC was 0.725 when comparing the child’s gold standard TV viewing time (min:sec/child) to FLASH-TV estimated time. The ICC for gaze/no gaze at the frame level (15-30 frames/sec) by FLASH-TV to the gold standard was 0.401.

Conclusion: FLASH-TV offers a critical step forward in improving the accuracy of assessment of children’s TV viewing time. A lower ICC at the frame level compared to aggregated time indicates reliability across families and smoothing between activities needs to be improved.  Once optimized, FLASH-TV can be used in surveillance and intervention studies to more precisely measure children’s TV viewing.

Purpose: Sedentary behaviors are considered the fourth greatest factor of mortality worldwide. Specifically, screen media use is the most popular leisure-time sedentary behavior among children and adolescents. It is, however, unknown how these sedentary behaviors relate to shapes of subcortical brain nuclei. We aimed to examine the association of sedentary behaviors (i.e., watching TV and playing video game, and total sedentary time) with the shapes of subcortical brain structures (i.e., hippocampus, accumbens, amygdala, caudate, pallidum, putamen, and thalamus) of children with overweight/obesity.

Methods: The present cross-sectional analyses used baseline data from 99 children with overweight/obesity (10.0  1.1 years; 60% boys) from the ActiveBrains project. Sedentary behavior was measured using the Youth Activity Profile-Spain (YAP-S) questionnaire, an adapted version of the original YAP. The shape of subcortical brain structures was assessed by magnetic resonance imaging, and its relationship with sedentary behaviors was examined after controlling for a set of potential confounders using a partial correlation permutation approach.

Results/findings: Our results showed that watching TV, playing video games and total sedentary time were selectively related to expansions and contractions of shapes of subcortical brain nuclei. Specifically, higher time spent watching TV was related to contractions in certain regions of the accumbens and pallidum (K ranging from 179 to 411; p < 0.05), while higher time spent playing video games was related to both expansions and contractions in certain regions of the hippocampus, caudate, pallidum, and thalamus (K ranging from 90 to 459; p < 0.05). Total sedentary time was mainly associated with contractions in caudate regions (K= 266). However, none sedentary behavior examined was associated with the amygdala (p > 0.05).

Conclusions: Sedentary behaviors may influence the shapes of subcortical brain structures in children with overweight/obesity. While watching TV and total sedentary time were related to contractions in basal ganglia subregions, video games were related to both expansions and contractions in certain subcortical regions. Future larger studies should confirm or contrast our findings, and shed light on its cognitive implications during childhood.