Except for adults age ≥60 yr, adolescents (age 12–18 yr) are the age group that spends the greatest amount of hours in sedentary behaviors (12). Within this age group, adolescent girls spend more time in sedentary behaviors than boys do. The potential health risk of sedentary behaviors has recently been acknowledged (15). The patterns by which the sedentary time is accrued may also play an important role in body fat accumulation and metabolic health. Three cross-sectional adult studies (1,4,5) have shown that the frequency of interruptions in sedentary time (hereafter “sedentary breaks”) is inversely associated with waist circumference, body mass index, C-reactive protein, fasting plasma glucose levels, and prevalence of the metabolic syndrome. However, to our knowledge, no study that examined the frequency of sedentary breaks in children and adolescents exists. In particular, longitudinal changes in the frequency of sedentary breaks have not been described. In this study, we describe the change in the frequency of sedentary breaks during a 10-yr period during childhood and early adolescence. We also compare the frequency between weekdays and weekend days. This study lays the foundation to begin to understand changes in patterns of sedentary behaviors during childhood and adolescence and to identify the potential effects of sedentary breaks on health outcomes among children and adolescents.
The Iowa Bone Development Study (IBDS) is an ongoing longitudinal study of bone health during childhood and adolescence. During 1998–2002, the IBDS recruited children at age 5 from the Iowa Fluoride Study population. The Iowa Fluoride Study originally recruited 1882 families of newborn babies from eight Iowa hospital postpartum wards from 1992 to 1995 (11). Approximately 95% of participants were white. Additional information about the study design and demographic characteristics of participants can be found in previous publications (6–9). The current analysis focused on data collected at approximately 5, 8, 11, 13, and 15 yr (labeled waves 1–5) during 1998–2009. Among children who were contacted and responded to a study invitation letter at wave 1, 455 of 478 consented to wear an accelerometer, as did 584 of 605 at wave 2, 553 of 581 at wave 3, 504 of 538 at wave 4, and 455 of 491 at wave 5. The study was approved by the University of Iowa Institutional Review Board (Human Subjects). Written informed consent was provided by the parents of the children. Assent was obtained from the children.
The detailed procedure for accelerometer data collection has been described in previous publications (6–9). During an initial study visit, research nurses trained in anthropometry measured the child’s height and weight. Children and their parents were also given instruction on accelerometer wear. ActiGraph uniaxial accelerometers (Model Number 7164; Pensacola, FL; GT1M for wave 5 only) were sent to eligible participants via prepaid mail during the fall season (September to November). Children were asked to wear the monitor during waking hours for four consecutive days, including one weekend day, at waves 1 and 2, and for five consecutive days, including both weekend days, at the other waves. At wave 5, a relatively higher proportion of participants wore the monitor for <4 d compared with previous waves. Therefore, those participants were asked to rewear the monitor for another 5 d, and data collection was extended until mid-December. Mid-December weather is similar to mid- to late-November weather in Iowa. Accelerometer movement counts were collected in 1-min epochs at waves 1–4 and 5-s epochs at wave 5. Wave 5 data were reintegrated to 1 min.
Accelerometer data reduction and processing
Accelerometers were considered as not worn if a period of 60 consecutive minutes of zero accelerometer counts with allowance for two nonzero interruptions was encountered in the accelerometer data array (3). We used only accelerometer data from participants who wore an accelerometer for a minimum of 10 h·d−1 and 3 d per wave (13). We identified 423, 550, 520, 454, and 344 individuals at waves 1–5, respectively, who met the inclusion criteria. Two hundred twenty-three participants had complete accelerometer data for all five waves: 174 for four waves, 101 for three waves, 62 for two waves, and 53 for one wave (a total of 613 individuals). Five individuals who wore the accelerometer again in December for wave 5 met the inclusion criteria.
Sedentary time was considered to be interrupted or broken if accelerometer counts were ≥100 counts per minute (4). The number of interruptions was counted as daily frequency of sedentary breaks (times per day). Moderate to vigorous intensity was defined as 2296 or greater accelerometer counts per minute (2,16). Accelerometer data were also organized by type of day (weekday and weekend day) and the following periods of day: monitor-on (morning) to 3:00 p.m., 3:00–7:00 p.m., and 7:00 p.m. to monitor-off (sleep).
Age was calculated by subtracting date of birth from date of accelerometer measurement. The hourly frequency of sedentary breaks (times per hour) was calculated by dividing daily frequency of sedentary breaks into daily wear time in hours for each wear day. Means of daily and hourly frequency of sedentary breaks per individual across the day of wear were calculated at each wave.
Descriptive analyses were conducted. Intraclass correlation coefficients (ICC) for repeated measures of the mean hourly frequency of sedentary breaks and moderate- to vigorous-intensity physical activity (MVPA) time were estimated in a multilevel model without predictors. Pearson correlation coefficients were estimated by sex and wave between the mean daily frequency of sedentary breaks and the mean daily MVPA time, as well as the mean daily sedentary time. To obtain aggregated correlation coefficients, the covariance correlation between the mean daily frequency of sedentary breaks and the mean daily MVPA time was calculated in a multilevel model.
Because a decreasing trend in the mean frequency of sedentary breaks over time was identified in the descriptive analysis, the yearly decrease in the mean hourly frequency of sedentary breaks was estimated using linear regression models. An age × sex interaction effect was significant (P < 0.05); therefore, sex-specific regression models were fitted (independent variable = age, dependent variable = the mean hourly frequency of sedentary breaks). To compare the hourly frequency of sedentary breaks between weekdays and weekend days, sex- and wave-specific multilevel models were built (fixed effect = age and type of the day, random effect = within-subject repeated measures for three or more days of monitor wear, dependent variable = the hourly frequency of sedentary breaks). Because between/within-interaction effects were not statistically significant, the interaction term was not included in the final models. Another set of multilevel models was built according to periods during the day. Matrix structure type was determined based on Akaike’s Information Criterion for goodness of fit. Residual and studentized residual graphs were used to confirm model assumptions and fit.
In exploratory work, we repeated analyses for a cohort of 223 individuals who had sedentary break data at all five waves (the “closed” cohort). In addition, we examined the tracking of frequency of sedentary breaks in the closed cohort using Pearson correlation coefficients. All analyses were conducted using SAS version 9.2 (Cary, NC).
Table 1 presents the characteristics of participants. The mean daily frequency of sedentary breaks decreased >200 times during a 10-yr period from age 5 to age 15 in both boys and girls. The mean daily sedentary time increased by approximately 4.5 h. The mean MVPA time was the highest at age 11 among boys, whereas it decreased over time among girls. The ICC for the mean hourly frequency of sedentary breaks was estimated as 0.06, whereas the ICC for the mean MVPA time was 0.39. Gender differences in the mean hourly frequency of sedentary breaks were within 2 times per hour at each wave, with fewer breaks among girls than among boys (P values < 0.01 at waves 3, 4, and 5). The slope from a fitted linear regression model estimated a 1.84-times-per-hour decrease in sedentary breaks per year for boys (R2 = 0.60) and a 2.04-times-per-hour decrease per year for girls (R2 = 0.66, P values < 0.0001).
As shown in Figure 1, there was a positive association between mean daily frequency of sedentary breaks and mean MVPA time (covariance correlation = 0.52). However, the mean daily frequency of sedentary breaks was widely distributed, particularly at low MVPA count levels. For each of waves 1–5, Pearson correlation coefficients of mean daily frequency of sedentary breaks and mean daily MVPA time were 0.41, 0.48, 0.60, 0.55, and 0.58, respectively, for boys, and 0.54, 0.43, 0.52, 0.46, and 0.50, respectively, for girls (all P values < 0.001). Wave-specific Pearson correlation coefficients of mean daily frequency of sedentary breaks and mean daily sedentary time were −0.71, −0.79, −0.74, −0.75, and −0.64, respectively, for boys, and −0.71, −0.78, −0.74, −0.64, and −0.55, respectively, for girls (all P values < 0.0001).
Figure 2 presents the mean hourly frequency of sedentary breaks over periods during the day by wave (age). In all waves, there was a pattern that the number of sedentary breaks was the lowest during morning to 3:00 p.m. and the highest from 3:00 to 7:00 p.m.
As illustrated in Figure 3A, boys and girls more frequently broke up their sedentary time on weekend days when compared to weekdays. When sorted by periods during the day, boys and girls had significantly fewer sedentary breaks on weekdays from morning to 3:00 p.m. compared to weekends from morning to 3:00 p.m. (Fig. 3B); the difference was six or more times per hour at all waves (P values < 0.0001), except for wave 1 when the difference was four times per hour (P values < 0.0001). From 3:00 to 7:00 p.m., the hourly frequency of sedentary breaks was not different between the weekdays and weekend days at waves 1–3 (Fig. 3C). However, both boys and girls had fewer sedentary breaks at waves 4 and 5 on weekend days when compared to weekdays (P values < 0.0001). After 7:00 p.m., both boys and girls had slightly fewer sedentary breaks on weekends when compared to weekdays (P values < 0.05; Fig. 3D).
The trend for a decreasing frequency of sedentary breaks over time was also observed in the closed cohort of 223 individuals (P values from linear regression estimates < 0.0001). Wave-specific means and SD of the mean hourly frequency of sedentary breaks were 43 ± 4, 38 ± 4, 35 ± 5, 32 ± 5, and 24 ± 4 times per hour, respectively, for 108 boys, and 42 ± 4, 37 ± 4, 34 ± 5, 29 ± 5, and 22 ± 4 times per hour, respectively, for 115 girls. As presented in Table 2, tracking of the mean hourly frequency of sedentary breaks for 2–3 yr was moderate (P values < 0.001).
This is the first study demonstrating a notable decrease in the frequency of sedentary breaks during childhood and adolescence. Boys and girls have fewer sedentary breaks during school hours than during non–school hours or weekend. Overall, gender differences in the frequency of sedentary breaks were minimal; although in general, girls took slightly fewer breaks than boys did. We found a moderate association between sedentary breaks and MVPA. However, the frequency of sedentary breaks varied widely for children with low levels of MVPA. The frequency of sedentary breaks was strongly and inversely associated with sedentary time. Finally, the tracking of the frequency of sedentary breaks for 2- to 3-yr intervals during childhood and early adolescence was moderate, whereas the tracking during a 10-yr interval was low.
Healy et al. (4) reported the first study to examine sedentary breaks among Australian adults, where the mean total number of sedentary breaks was 601 times during the mean total sedentary time of 56.7 h. In another study by Healy et al. (5), the average frequency of daily sedentary breaks was 90 times per day among men (mean wear time = 14.9 h·d−1) and 95 times per day among women (mean wear time = 14.4 h·d−1) participating in the 2003–2006 NHANES. It seems that adults have considerably fewer sedentary breaks than the IBDS cohort members do. This comparison may reflect the different social circumstances and lifestyle choices throughout the life cycle. Children’s activity patterns are more sporadic and spontaneous. As they enter adolescence, they are involved in longer sitting activities in the classroom and during leisure time (e.g., computer use). The lifestyle of adults often involves continuous and prolonged sedentary behaviors at work and during leisure time. Gender differences seem to be minimal, although they should be further verified by data from diverse populations.
Repeated measures of the frequency of sedentary breaks during a 10-yr period were weakly correlated when compared to repeated measures of MVPA time. This finding suggests that sedentary break behaviors may be less stable than MVPA behaviors during childhood and early adolescence. The tracking of sedentary breaks during a 10-yr period was low. This is advantageous for intervention programs aimed at decreasing prolonged inactivity because our findings suggest a behavior that is changeable.
The frequency of sedentary breaks was strongly and inversely associated with sedentary time. This implies an increased engagement in prolonged sedentary behaviors, such as TV viewing or computer use, as children grow into adolescence. There are increased expectations within the schools that children spend prolonged time on tasks as they advance through grades. Interventions that simply focus on breaking prolonged sedentary time might be successful in reducing total sedentary time. A great deal of variability in the frequency of sedentary breaks within a low level of MVPA suggests the potential independence of MVPA and sedentary break behaviors among those who have low MVPA levels. This finding indirectly suggests that light activity such as sedentary breaks may be particularly important and effective to adolescents who engage in low MVPA.
Although we found that sedentary breaks were significantly fewer (approximately six times per hour) on weekdays from morning to 3:00 p.m. (school hours) than weekend days from morning to 3:00 p.m., no evidence exists on whether the magnitude of the difference is meaningful with respect to health risk. Future research is recommended to investigate the effects of sedentary breaks on various health outcomes, particularly adiposity, among adolescents. If health effects of sedentary breaks are proven, schools may be a potential intervention setting for increasing the frequency of sedentary breaks. School policy interventions such as increasing recess time and activity breaks during class time (14) may have an effect on the students’ sedentary break behaviors. Interventions in the school setting may provide additional benefits in academic performance and morale.
The main limitation of this study is the threat of selection bias. We included participants who completed at least one of five assessments (open cohort). Loss to follow-up may have caused selection bias. Our cohort study experienced significant attrition particularly at age 15. However, analyses for the closed cohort revealed that the patterns of sedentary breaks were similar to the open cohort. Sedentary break behaviors may be influenced by maturation and sociopsychological changes during adolescence, which the current study did not take into account. The frequency in sedentary breaks may differ by season, and therefore, our data collected in the fall season may not represent the mean annual frequency. The frequency in sedentary breaks may differ by season, and therefore, our data collected in the fall season may not represent the mean annual frequency. However, the repeated measurements during the same season would have minimized bias by seasonal variation within a year. This is particularly important given our cohort is mainly in the Midwest which has dramatically different weather patterns across seasons. The accelerometry output may differ between ActiGraph models 7164 and GT1M: greater variability in model 7164 with analog circuit filtering system than GT1M with digital filtering system. We calibrated all the model 7164 accelerometers with a shaker before sending out to participants. Also, it has been reported that the difference is random, and data with the GM1 are comparable with those from model 7164 when summarizing data using cut point methods (10). Owing to the geographical location of the IBDS, most participants are white. Therefore, caution should be taken in generalizing the results to diverse populations in other geographical locations.
Nonetheless, this study is innovative as a first longitudinal investigation of sedentary breaks during childhood and adolescence. Five assessments during a 10-yr period allowed demonstration of the change in sedentary breaks throughout childhood and early adolescence. This study would provide a new venue for research on the roles of sedentary behaviors in childhood obesity and health.
In conclusion, this study demonstrates that the frequency of sedentary breaks notably decreases over childhood and adolescence. Boys and girls have fewer sedentary breaks during school hours than during after-school hours or during the weekends.
This study was supported by the National Institute of Dental and Craniofacial Research (R01-DE12101 and R01-DE09551) and the General Clinical Research Centers Program from the National Center for Research Resources (M01-RR00059).
The authors have no conflicts of interests to report.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Bankoski A, Harris TB, McClain JJ, et al.. Sedentary activity associated with metabolic syndrome independent of physical activity. Diabetes Care. 2011; 34 (2): 497–503.
2. Evenson KR, Catellier DJ, Gill K, Ondrak KS, McMurray RG. Calibration of two objective measures of physical activity for children. J Sports Sci. 2008; 26 (14): 1557–65.
3. Evenson KR, Terry JW Jr. Assessment of differing definitions of accelerometer nonwear time. Res Q Exerc Sport. 2009; 80 (2): 355–62.
4. Healy GN, Dunstan DW, Salmon J, et al.. Breaks in sedentary time: beneficial associations with metabolic risk. Diabetes Care. 2008; 31 (4): 661–6.
5. Healy GN, Matthews CE, Dunstan DW, Winkler EA, Owen N. Sedentary time and cardio-metabolic biomarkers in US adults: NHANES 2003–06. Eur Heart J. 2011; 32 (5): 590–7.
6. Janz KF, Burns TL, Levy SM, Iowa Bone Development Study. Tracking of activity and sedentary behaviors in childhood: the Iowa Bone Development Study. Am J Prev Med. 2005; 29 (3): 171–8.
7. Janz KF, Burns TL, Torner JC, et al.. Physical activity and bone measures in young children: the Iowa Bone Development Study. Pediatrics. 2001; 107 (6): 1387–93.
8. Janz KF, Kwon S, Letuchy EM, et al.. Sustained effect of early physical activity on body fat mass in older children. Am J Prev Med. 2009; 37 (1): 35–40.
9. Janz KF, Levy SM, Burns TL, Torner JC, Willing MC, Warren JJ. Fatness, physical activity, and television viewing in children during the adiposity rebound period: the Iowa Bone Development Study. Prev Med. 2002; 35 (6): 563–71.
10. Kozey SL, Staudenmayer JW, Troiano RP, Freedson PS. Comparison of the ActiGraph 7164 and the ActiGraph GT1M during self-paced locomotion. Med Sci Sports Exerc. 2010; 42 (5): 971–6.
11. Levy SM, Warren JJ, Davis CS, Kirchner HL, Kanellis MJ, Wefel JS. Patterns of fluoride intake from birth to 36 months. J Public Health Dent. 2001; 61 (2): 70–7.
12. Matthews CE, Chen KY, Freedson PS, et al.. Amount of time spent in sedentary behaviors in the United States, 2003–2004. Am J Epidemiol. 2008; 167 (7): 875–81.
13. Mattocks C, Ness A, Leary S, et al.. Use of accelerometers in a large field-based study of children: protocols, design issues, and effects on precision. J Phys Act Health. 2008; 5 (Suppl 1): S98–111.
14. Salmon J. Novel strategies to promote children’s physical activities and reduce sedentary behavior. J Phys Act Health. 2010; 7 (Suppl 3): S299–306.
15. Salmon J, Tremblay MS, Marshall SJ, Hume C. Health risks, correlates, and interventions to reduce sedentary behavior in young people. Am J Prev Med. 2011; 41 (2): 197–206.
16. Trost SG, Loprinzi PD, Moore R, Pfeiffer KA. Comparison of accelerometer cut-points for predicting activity intensity in youth. Med Sci Sports Exerc. 2011; 43 (7): 1360–8.