Participants
Children (8–13 years) and adolescents (14–18 years) were recruited from 11 elementary and secondary schools. Schools with a specific focus on sport and schools for pupils with special educational needs were not included. Participants were recruited to participate on a voluntary basis via information flyers that were distributed through the school staff after the school management approved the research. The main inclusion criteria were participant age and good health condition. The participants whose parents reported medical complications that could affect PA and sleep were excluded from study. A total of 907 children and adolescents were enrolled in this study. Of all initial participants, 228 were excluded because they voluntarily withdrew from the study or became ill (n = 45), provided incomplete data (n = 129), their data could not be assessed due technical failures (n = 17), or did not meet accelerometer wear time criteria (n = 37). Hence, the final sample consisted of 355 children (56% girls) and 324 adolescents (57% girls). The detailed characteristics of the participants are shown in Table 1.
Physical activity and sleep
The amount of time spent in MVPA and sleep duration were estimated using the wGT3X-BT and GT9X Link ActiGraph accelerometers (ActiGraph, Pensacola, FL, USA) worn by children and adolescents, respectively. Participants wore the activity monitor on their non-dominant wrist for 24 hours over 7 consecutive days. They were instructed to remove the device only for swimming and bathing. Participants recorded their times of wake up, falling asleep, and non-wear periods in a daily log. The device was initialized using the ActiLife software version 6.13.3 (ActiGraph, Pensacola, FL, USA), all three axes were used, and sampling interval was set to 100 Hz. To limit reactivity, the displays of GT9X Link accelerometers were set to show only date and time. Raw accelerometer data were analyzed using the R-package GGIR version 1.10-7. A more detailed description has been published elsewhere [39].
Screen time
Recreational ST was self-reported. A parent proxy report was required in children aged 12 years and younger (i.e., those in the first stage of elementary school). Participants, their parents, or guardians answered the questions taken from the questionnaire of the international Health Behaviour in School-aged Children study [40] as follows: “About how many hours a day do you usually spend watching television, DVDs, videos (including YouTube or similar online service) in your free time on weekdays/weekend days?” and “About how many hours a day do you usually spend playing games on a computer, games console (PlayStation, Xbox, etc.), smartphone, tablet or similar electronic device in your free time on weekdays/weekend days?”. Questions were separated for weekdays and weekend days. Nine different answers were available for each question (none, half an hour, 1, 2, 3, 4, 5, 6, and 7 or more hours a day). The validity and reliability of 7-day recall questions have been demonstrated in comparison with 7-day 24-hour diaries both on weekdays and weekends [41]. Total amount of ST was calculated as the sum of weighted averages of ST during weekdays and weekend days.
Adherence to the combined movement guidelines
Participants adhere to the combined movement guidelines if they accumulate at least 60 minutes of MVPA per day for PA recommendation, 2 hours or less of recreational ST per day for SB recommendation, and 9–11 hours per day for children and 8–10 hours per day for adolescents for sleep recommendation.
Correlates
Sixteen potential correlates were selected based on systematic reviews [2,34–38] showing plausible associations with at least single recommendation included in the combined movement guidelines. Correlates were grouped into three categories: (1) biological and cognitive correlates, (2) behavioral correlates, and (3) family correlates. They were obtained through multiple research sources. Biological correlates were measured directly using standard anthropometric measurements and the multi-frequency bioimpedance analyzer InBody 720 (InBody, Seoul, Korea). Cognitive and behavioral correlates were self-reported except for sleep efficiency, which was measured by accelerometry. Parent proxy report was required for participants aged 12 years and younger. Family correlates were reported by parents. The full list of correlates with information about their use in the analysis is displayed in Table S1.
Procedure
Data were collected from 2018 to 2019 during regular school weeks. Participants were given accelerometers in the classrooms and were instructed on how to wear them properly and how to complete relevant daily logs. Although the participants could already use the devices, the official start of monitoring was set for the next full day to minimize reactivity bias. Participants and their parents or guardians were asked to fill in the questionnaires.
Statistical analyses
Statistical analyses were conducted using the IBM SPSS Statistics version 23 (IBM, Armonk, NY, USA) and R version 3.4.2 (R Foundation for Statistical Computing, Vienna, Austria). The differences between children and adolescents were analyzed using the t-test for continuous variables and the chi-squared test for categorical variables.
Univariable analysis was conducted to examine associations between potential correlates and adherence to the combined movement guidelines and the specific combinations of any two recommendations. Binary logistic regression models were used because of the inherent nature of dependent variables ("0" for not meeting and "1" for meeting the combined movement guidelines or combinations of any two recommendations). If an explanatory variable reached a less-strict criterion level of p < 0.1, it was retained for further analysis to prevent the exclusion of potentially important correlates.
Multi-level multivariable analysis was performed to identify correlates of adhering to the combined movement guidelines and of meeting combinations of any two recommendations. The potential correlates and sex of participants were included in the final models as fixed effects (Level 1), while the school location was considered a random effect (Level 2) in all mixed effects models. The necessity to include the factor of school location in the model was tested (by the likelihood-ratio test) and the factor was omitted whenever possible. Odds ratios (OR) and the 95% confidence intervals (CI) corresponding to the individual correlates as well as their significance were calculated. The forward selection method was used to set up the final model. The final models include all correlates whose omission would lead to a significant decrease in the Akaike information criterion. All statistical analyses were conducted at a significance level of p < 0.05.
Table 1
Descriptive characteristics of children and adolescents
| | Children n = 355 | | Adolescents n = 324 | | p-valueb |
| | Mean | SD | | Mean | SD | | |
Personal data | | | | | | | | |
Age (years) | | 11.7 | 1.6 | | 16.3 | 1.3 | | < 0.001 |
Height (cm) | | 151.6 | 12.0 | | 170.2 | 8.8 | | < 0.001 |
Weight (kg) | | 43.6 | 11.3 | | 63.0 | 11.6 | | < 0.001 |
BMI z-score | | 0.24 | 1.13 | | 0.20 | 0.99 | | 0.587 |
Movement behaviors | | | | | | | | |
MVPA (min/day)a | | 58.1 | 24.3 | | 39.3 | 19.1 | | < 0.001 |
ST (h/day) | | 3.0 | 1.8 | | 2.8 | 2.1 | | 0.206 |
Sleep duration (h/day)a | | 8.6 | 0.7 | | 7.5 | 0.8 | | < 0.001 |
BMI: Body mass index; MVPA: Moderate-to-vigorous physical activity; ST: Screen time; SD: Standard deviation |
a Accelerometer-based 24-hour assessment; adjusted to 24 hours before analysis |
b The differences between age categories were analyzed using the t-test for independent samples |