[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.205.0.26. Please contact the publisher to request reinstatement.
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure 1.
Mean MVPA Minutes and 95% CIs by MVPA Trajectory Classes
Mean MVPA Minutes and 95% CIs by MVPA Trajectory Classes

Dots indicate actual mean moderate- to vigorous-intensity physical activity (MVPA) minutes; solid lines, estimated mean MVPA minutes; and dotted lines, 95% CIs of estimated mean MVPA minutes.

Figure 2.
Probability of Organized Sports Participation and 95% CIs by Organized Sports Participation Trajectory Classes
Probability of Organized Sports Participation and 95% CIs by Organized Sports Participation Trajectory Classes

Dots indicate actual probability of sports participation; solid lines, estimated probability of sports participation; and dotted lines, 95% CIs of estimated probability of sports participation.

Figure 3.
Mean Television Viewing Hours and 95% CIs by Television Viewing Trajectory Classes
Mean Television Viewing Hours and 95% CIs by Television Viewing Trajectory Classes

Dots indicate actual mean television viewing hours; solid lines, estimated mean television viewing hours; and dotted lines, 95% CIs of estimated mean television viewing hours.

Table 1.  
Descriptive Analysis of Physical Activity Measures Across 7 Examinations of the Iowa Bone Development Study
Descriptive Analysis of Physical Activity Measures Across 7 Examinations of the Iowa Bone Development Study
Table 2.  
Distribution of MVPA Trajectory Groups Conditional on Organized Sports Participation Trajectory Groups and on Television Viewing Trajectory Groups
Distribution of MVPA Trajectory Groups Conditional on Organized Sports Participation Trajectory Groups and on Television Viewing Trajectory Groups
1.
Eime  RM, Young  JA, Harvey  JT, Charity  MJ, Payne  WR.  A systematic review of the psychological and social benefits of participation in sport for children and adolescents: informing development of a conceptual model of health through sport. Int J Behav Nutr Phys Act. 2013;10:98.
PubMedArticle
2.
Physical Activity Guidelines Advisory Committee. Physical Activity Guidelines Advisory Committee Report. Washington, DC: US Dept of Health and Human Services; 2008.
3.
Floriani  V, Kennedy  C.  Promotion of physical activity in children. Curr Opin Pediatr. 2008;20(1):90-95.
PubMedArticle
4.
Troiano  RP, Berrigan  D, Dodd  KW, Mâsse  LC, Tilert  T, McDowell  M.  Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181-188.
PubMedArticle
5.
US Department of Health and Human Services. Promoting Better Health for Young People Through Physical Activity and Sports. Atlanta, GA: The Centers for Disease Control and Prevention; 2000.
6.
Geidne  S, Quennerstedt  M, Eriksson  C.  The youth sports club as a health-promoting setting: an integrative review of research. Scand J Public Health. 2013;41(3):269-283.
PubMedArticle
7.
Spinks  A, Macpherson  A, Bain  C, McClure  R.  Determinants of sufficient daily activity in Australian primary school children. J Paediatr Child Health. 2006;42(11):674-679.
PubMedArticle
8.
Nilsson  A, Andersen  LB, Ommundsen  Y,  et al.  Correlates of objectively assessed physical activity and sedentary time in children: a cross-sectional study (The European Youth Heart Study). BMC Public Health. 2009;9:322.
PubMedArticle
9.
Wickel  EE, Eisenmann  JC.  Contribution of youth sport to total daily physical activity among 6- to 12-yr-old boys. Med Sci Sports Exerc. 2007;39(9):1493-1500.
PubMedArticle
10.
Katzmarzyk  PT, Malina  RM.  Contribution of organized sports participation to estimated daily energy expenditure in youth. Pediatr Exerc Sci. 1998;10(4):378-386.
11.
Sallis  JF, Prochaska  JJ, Taylor  WC.  A review of correlates of physical activity of children and adolescents. Med Sci Sports Exerc. 2000;32(5):963-975.
PubMedArticle
12.
Nelson  TF, Stovitz  SD, Thomas  M, LaVoi  NM, Bauer  KW, Neumark-Sztainer  D.  Do youth sports prevent pediatric obesity? a systematic review and commentary. Curr Sports Med Rep. 2011;10(6):360-370.
PubMedArticle
13.
Sabo  D, Veliz  P. Go Out and Play: Youth Sports in America. East Meadow, NY: Women's Sports Foundation; 2008.
14.
Robinson  TN.  Television viewing and childhood obesity. Pediatr Clin North Am. 2001;48(4):1017-1025.
PubMedArticle
15.
Gebremariam  MK, Bergh  IH, Andersen  LF,  et al.  Are screen-based sedentary behaviors longitudinally associated with dietary behaviors and leisure-time physical activity in the transition into adolescence? Int J Behav Nutr Phys Act. 2013;10:9.
PubMedArticle
16.
Motl  RW, McAuley  E, Birnbaum  AS, Lytle  LA.  Naturally occurring changes in time spent watching television are inversely related to frequency of physical activity during early adolescence. J Adolesc. 2006;29(1):19-32.
PubMedArticle
17.
Koezuka  N, Koo  M, Allison  KR,  et al.  The relationship between sedentary activities and physical inactivity among adolescents: results from the Canadian Community Health Survey. J Adolesc Health. 2006;39(4):515-522.
PubMedArticle
18.
Epstein  LH, Roemmich  JN, Paluch  RA, Raynor  HA.  Physical activity as a substitute for sedentary behavior in youth. Ann Behav Med. 2005;29(3):200-209.
PubMedArticle
19.
Robinson  TN, Hammer  LD, Killen  JD,  et al.  Does television viewing increase obesity and reduce physical activity? cross-sectional and longitudinal analyses among adolescent girls. Pediatrics. 1993;91(2):273-280.
PubMed
20.
Taveras  EM, Field  AE, Berkey  CS,  et al.  Longitudinal relationship between television viewing and leisure-time physical activity during adolescence. Pediatrics. 2007;119(2):e314-e319.
PubMedArticle
21.
Melkevik  O, Torsheim  T, Iannotti  RJ, Wold  B.  Is spending time in screen-based sedentary behaviors associated with less physical activity: a cross national investigation. Int J Behav Nutr Phys Act. 2010;7:46.
PubMedArticle
22.
Nagin  D. Group-Based Modeling of Development. Cambridge, MA: Harvard University Press; 2005.
23.
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):70-77.
PubMedArticle
24.
Janz  KF, Burns  TL, Levy  SM.  Tracking of activity and sedentary behaviors in childhood: the Iowa Bone Development Study. Am J Prev Med. 2005;29(3):171-178.
PubMedArticle
25.
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.
PubMedArticle
26.
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-S111.
PubMed
27.
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):1557-1565.
PubMedArticle
28.
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):1360-1368.
PubMedArticle
29.
Kowlaski  KC, Crocker  PRE, Faulkner  RA.  Validation of the Physical Activity Questionnaire for older children. Pediatr Exerc Sci. 1997;9(2):174-186.
30.
Kowlaski  KC, Crocker  PRE, Kowlaski  NP.  Convergent validity of the Physical Activity Questionnaire for adolescents. Pediatr Exerc Sci. 1997;9(4):342-352.
31.
Jones  B, Nagin  D. A Stata plugin for estimating group-based trajectory models.2012. http://ssrc.indiana.edu/seminars/wimdocs/2013-03-29_nagin_trajectory_stata-plugin-info.pdf. Accessed March 26, 2015.
32.
Nylund  KL, Asparouhov  T, Muthen  BO.  Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Modeling. 2007;14(4):535-569. doi:10.1080/10705510701575396.Article
33.
Janz  KF, Letuchy  EM, Burns  TL, Eichenberger Gilmore  JM, Torner  JC, Levy  SM.  Objectively measured physical activity trajectories predict adolescent bone strength: Iowa Bone Development Study. Br J Sports Med. 2014;48(13):1032-1036.
PubMedArticle
34.
Troiano  RP, McClain  JJ, Brychta  RJ, Chen  KY.  Evolution of accelerometer methods for physical activity research. Br J Sports Med. 2014;48(13):1019-1023.
PubMedArticle
35.
Slootmaker  SM, Schuit  AJ, Chinapaw  MJ, Seidell  JC, van Mechelen  W.  Disagreement in physical activity assessed by accelerometer and self-report in subgroups of age, gender, education and weight status. Int J Behav Nutr Phys Act. 2009;6:17.
PubMedArticle
36.
Vella  SA, Cliff  DP, Okely  AD.  Socio-ecological predictors of participation and dropout in organised sports during childhood. Int J Behav Nutr Phys Act. 2014;11:62.
PubMedArticle
37.
Walters  S, Barr-Anderson  DJ, Wall  M, Neumark-Sztainer  D.  Does participation in organized sports predict future physical activity for adolescents from diverse economic backgrounds? J Adolesc Health. 2009;44(3):268-274.
PubMedArticle
38.
Sabo  D, Veliz  P. Progress Without Equity: The Provision of High School Athletic Opportunity in the United States, by Gender 1993-94 Through 2005-2006. New York, NY: Women's Sports Foundation; 2011.
39.
Leech  RM, McNaughton  SA, Timperio  A.  The clustering of diet, physical activity and sedentary behavior in children and adolescents: a review. Int J Behav Nutr Phys Act. 2014;11:4.
PubMedArticle
40.
Roeder  K, Lynch  K, Nagin  D.  Modeling uncertainty in latent class membership: a case study in criminology. J Am Stat Assoc. 1999;94(447):766-776.Article
Original Investigation
July 2015

Developmental Trajectories of Physical Activity, Sports, and Television Viewing During Childhood to Young AdulthoodIowa Bone Development Study

Author Affiliations
  • 1Stanley Manne Children’s Research Institute, Ann & Robert H. Lurie Children’s Hospital of Chicago, Chicago, Illinois
  • 2Department of Health and Human Physiology, University of Iowa, Iowa City
  • 3Department of Epidemiology, University of Iowa, Iowa City
  • 4Department of Preventive and Community Dentistry, University of Iowa, Iowa City
JAMA Pediatr. 2015;169(7):666-672. doi:10.1001/jamapediatrics.2015.0327
Abstract

Importance  The diverse developmental patterns of obesogenic behaviors during childhood and adolescence can be better understood by using new analytic approaches to assess the heterogeneity in variation during growth and development and to map the clustering of behavior patterns.

Objectives  To identify distinct trajectories of daily time spent in moderate- to vigorous-intensity physical activity (MVPA) from ages 5 to 19 years and to examine the associations of MVPA trajectories with sports participation and television viewing trajectories.

Design, Setting, and Participants  Cohort members in the prospective population-based Iowa Bone Development Study participated in MVPA assessments via accelerometry from September 16, 1998, to December 9, 2013, at ages 5, 8, 11, 13, 15, 17, and 19 years and completed a questionnaire every 6 months on sports participation and daily time spent in television viewing.

Main Outcomes and Measures  Trajectories of MVPA (minutes per day), participation in organized sports (yes or no), and television viewing time (hours per day).

Results  Based on the data from 537 participants (50.1% females; 94.6% white), we identified 4 MVPA trajectories: consistently inactive (14.9%), consistently active (18.1%), decreasing moderate physical activity (52.9%), and substantially decreasing high physical activity (14.1%). All participants in the consistently inactive trajectory also followed a trajectory of no participation in sports. The consistently active trajectory was associated with decreasing an already low television viewing trajectory (P < .001).

Conclusions and Relevance  This study provided a nuanced look at the known decrease in MVPA during childhood and adolescence. Sports participation could be a critical way to avoid the consistently inactive pattern. Most important, we identified a subset of participants who maintained a seemingly healthy level of MVPA from childhood to young adulthood. The developmental pathways of physical activity and television viewing behaviors could be related. Additional studies should examine the determinants and health consequences of these specific MVPA trajectories.

Introduction

Participation in moderate- to vigorous-intensity physical activity (MVPA) during childhood provides numerous physical, psychosocial, and cognitive health benefits.13 However, most US youth participate in less MVPA than the recommended 60 min/d.4 Organized sports participation, as a leisure-time activity, has been promoted as an effective strategy to increase physical activity (PA) among youth.5,6 Because sports is a subset of PA, it is not surprising to find a strong cross-sectional association between organized sports participation and MVPA among youth.711 Longitudinal studies have also reported a positive association between organized sports participation and PA during adolescence.12 These data suggest that consistent participation in organized sports during the growing years is associated with healthy MVPA habits. In reality, of the 84% of US youth who join organized sports, 45% will stop participation at some point during childhood and adolescence.13 Therefore, it is important to understand whether any participation in organized sports helps youth to develop healthier MVPA habits compared with never participating in sports at all.

Television viewing has been seen as a displacement behavior for PA; that is, youth watch television rather than engage in PA.14 However, it also has been viewed as an independent health risk behavior that is not necessarily associated with PA levels. In fact, the association between television viewing and PA is mixed: while some studies1517 report an inverse cross-sectional and longitudinal association, other studies report no association,1820 particularly in populations with relatively low levels of PA.21 A better understanding of the developmental pathways of PA and television viewing behaviors will provide greater insight for improving interventions that target both or each of these behaviors.

Identifying distinct developmental patterns of MVPA, sports, and television viewing behaviors requires an analytic method that effectively uses multiple points of data assessed longitudinally. Traditional analytic approaches generally examine the paths of predefined subgroups using prior analysis and subjective classification rules; that is, groups with an increase, decrease, or no change are classified based on a change between 2 time points. Such predefined subgroup approaches are fraught with statistical dangers, including the dual risks of creating groups that reflect only random variation and failing to identify important but unusual developmental patterns.22 Group-based trajectory modeling enables the data themselves to divide a study population into subgroups and to fit trajectory models for each of those subgroups.22 The aims of this study were to use group-based modeling to identify distinct developmental trajectories of MVPA, organized sports participation, and television viewing behaviors from childhood to young adulthood in the Iowa Bone Development Study (IBDS) cohort and to examine associations of the MVPA trajectories with organized sports participation trajectories and television viewing trajectories.

Box Section Ref ID

At a Glance

  • • This group-based trajectory approach offers an understanding of the development of obesogenic behaviors in youth.

  • • Group-based trajectory analysis identified a subset of about 20% of participants who maintained a seemingly healthy level of moderate- to vigorous-intensity physical activity from childhood to young adulthood and watched even less television across time.

  • • The developmental pathways of healthy behaviors (high physical activity and low television viewing) could be related.

  • • Sport participation could be a critical way to avoid a consistently inactive pattern.

Methods
Participants

The IBDS is an ongoing longitudinal study of bone health during childhood, adolescence, and young adulthood. Participants are a subset of the Iowa Fluoride Study birth cohort that included 1882 families recruited from 8 Iowa hospital postpartum wards between 1993 and 1997.23 The IBDS used rolling admission to allow Iowa Fluoride Study members to participate in any follow-up examinations. The baseline examinations for the IBDS were conducted between September 16, 1998, and October 5, 2000, in children who were approximately 5 years of age. Participants’ parents self-identified their race among the categories of white, black or African-American, Hispanic or Latino, Asian or Pacific Islander, and other. Of the 630 IBDS participants who completed at least 1 accelerometry assessment wave, 94.0% were white and 72.2% of their mothers had a college degree or higher educational level. Additional information about the IBDS and its participants is available elsewhere.24,25 Written informed consent and assent were obtained. The University of Iowa Institutional Review Board approved the IBDS, and the Ann & Robert H. Lurie Children’s Hospital of Chicago Institutional Review Board approved a waiver from review.

Measurements
Accelerometry-Measured PA

Accelerometry assessments were conducted at approximately 5, 8, 11, 13, 15, 17, and 19 years of age (IBDS waves 1 to 7, respectively) from September 16, 1998, to December 9, 2013. When children and their parents visited study sites for onsite examinations, they were given instructions for wearing the accelerometer. ActiGraph uniaxial accelerometers (ActiGraph; model 7164 for waves 1-4, GT1M for wave 5, and GT3X+ for waves 6 and 7) were mailed to participants during the autumn (September-November). The detailed procedures for accelerometer data collection are described in previous publications.24,25 Briefly, participants were asked to wear the monitors during waking hours for 4 consecutive days (including 1 weekend day) during waves 1 and 2 and for 5 consecutive days (including both weekend days) for all other waves. Accelerometry movement counts were collected in 1-minute epochs during waves 1 to 4 and in 5-second epochs during waves 5 to 7. For analysis, the 5-second epoch raw data were reintegrated into 1-minute epoch data for consistency across the waves. For the accelerometry data reduction process, accelerometers were considered as having not been worn if 60 consecutive minutes of zero accelerometry counts (with an allowance for 2 nonzero interruptions) was encountered in the accelerometry data array. Accelerometry data were used only from participants who wore an accelerometer for at least 8 h/d and at least 3 days during each wave.26 Daily wear time began at 12:00 am and ended at 11:59 pm; MVPA (as minutes per day) was defined as 2296 or more accelerometry counts per minute.27,28

Organized Sports Participation and Television Viewing

To define organized sports, we included organized teams and individual sports in any setting (eg, school-based sports teams and sports clubs) that were plausible in the Midwestern United States (as judged by 3 of us: S.K., E.M.L., and K.F.J.). Eligible sports included baseball, softball, basketball, cheerleading, dance (ballet, jazz, or other), football, golf, gymnastics, tumbling, ice hockey, figure skating, martial arts, track and field running, soccer, swimming, tennis, badminton, racquetball, volleyball, lacrosse, and wrestling. To assess organized sports participation and television viewing, 2 versions of a physical activity questionnaire (PAQ) were used. The first version (proxy-PAQ) was designed for a parent to report his or her child’s activities and was used for children younger than 11 years (waves 1 and 2). The proxy-PAQ asked about sports participation: “Did your child participate in any of the following organized sports during the past 6 months?” Responses included baseball, basketball, soccer, gymnastics, dance, swimming, and other. Responses of “other” were reviewed to capture any of the eligible sports listed as “other.” Children who were reported as having participated in at least 1 eligible sport were categorized as participating in organized sports. The proxy-PAQ also asked an open-ended question about the hours per day the child spent watching any type of television (including videotapes of movies and programs).

The second version (self-PAQ), administered when participants were aged 11 years or older, was adapted from the PAQ for older children29 and PAQ for adolescents.30 The self-PAQ contained a list of 20 or more sports, and inquired about the frequency of sports participation in the previous 7 days and whether the sport activity was organized, described as a coached sport or lessons. Responses of “other” were also allowed and reviewed by investigators to capture any of the eligible sports listed as “other.” If a respondent reported participating in at least 1 eligible sport and marked it as organized, the respondent was categorized as participating in organized sports. The self-PAQ asked, “How many hours per day did you watch television in the last week? Include the time you spent watching movies or programs on video or DVD. Do not count television or videos watched in school.” The response options of less than 1, 1 to less than 2, 2 to less than 3, 3 to less than 4, and 4 h/d or greater were coded as 0.5, 1.5, 2.5, 3.5, and 4.5, respectively, and were treated as continuous data. All PAQs were administered using a paper form either during clinic visits or via mail every 6 months during the individual-level 14-year follow-up.

Statistical Analysis

Participants who completed at least 3 of the 7 accelerometry assessments were included in the data analyses. This approach allowed us to consider the possibility of quadratic models. Descriptive analyses for the main variables were performed by waves (at approximately 5, 8, 11, 13, 15, 17, and 19 years of age). To identify distinct MVPA trajectory patterns (groups), we conducted group-based trajectory analysis with STATA TRAJ (StataCorp LP).31 In the process of determining the number of groups, a quadratic model was initially used for all groups. The final number of groups was determined based on the Bayesian Information Criterion, trajectory shapes for similarity, and the proportion of cohort members in each class.32 After identifying the optimal number of groups, the level of polynomial functions for each group (ie, quadratic, linear, and constant) was reduced until a parameter estimate in the highest function had a significance of P < .01 (eTable 1 in the Supplement). As suggested by Nagin,22 we evaluated model adequacy using the following 4 diagnostic measures: average posterior probability of assignment for each group is 0.7 or higher, odds of correct classification are 5.0 or higher, the proportion of a sample assigned to a certain group is close to the proportion estimated from the model, and 99% CIs of the estimated proportion are reasonably narrow. This modeling process was repeated for organized sports participation (yes or no) and television viewing (hours per day). We then used dual trajectory models of MVPA and organized sports participation, as well as MVPA and television viewing, to understand the interrelationship between the development of MVPA and television viewing behaviors because the dual trajectory model offers probabilities that link membership in trajectory groups across 2 behaviors.22 Each participant was assigned to one of the trajectory groups based on the individual’s maximum posterior probability from the final models. The difference in the sex distribution between the groups was tested via χ2 tests. P < .05 was considered significant.

Results

Of 630 participants who completed at least 1 accelerometry assessment wave, 537 participants (50.1% females) who had at least 3 valid accelerometer assessment waves were included in the data analysis. Those who were excluded (n = 93) had similar levels of family income and mother’s education as those who were included (eTable 2 in the Supplement). Of the 537 participants, 31.3% completed all 7 MVPA assessments, 29.0% completed 6 assessments, 16.4% completed 5 assessments, 14.0% completed 4 assessments, and 9.3% completed 3 assessments. The number of accelerometry assessments completed at each wave is presented in Table 1. The mean accelerometry wear time at each wave was 12 hours or more (range, 12.0-12.7 hours across waves). The median of accelerometry wear days was 5 in all waves. On average, each participant completed the PAQ 23 times during the study years (range, 3-35 times). In general, more participants completed the PAQ than the accelerometry assessment at each wave (Table 1).

We identified 4 distinct MVPA trajectories (Figure 1 and eTable 3 in the Supplement): consistently inactive (MVPA group 1; 14.9%), consistently active (MVPA group 2; 18.1%), decreasing moderate PA (MVPA group 3; 52.9%), and substantially decreasing high PA (MVPA group 4; 14.1%). Compared with males, more females followed the consistently inactive pattern (P < .001), but fewer females followed the consistently active pattern (P < .001). The percentages of females were 84.8% in MVPA group 1, 28.6% in MVPA group 2, 57.8% in MVPA group 3, and 11.8% in MVPA group 4.

We identified 3 organized sports participation trajectories (Figure 2): no sport participation (sport group 1; 13.6%), dropout from sports participation (sport group 2; 40.2%), and consistent sports participation (sport group 3; 46.2%). The sex distributions were not significantly different by groups (P = .88): the percentages of females were 52.9% in sport group 1, 49.5% in sport group 2, and 49.8% in sport group 3.

We identified 4 television viewing trajectories (Figure 3): decreasing low television viewing (television group 1; 25.4%), increasing low television viewing (television group 2; 26.4%), decreasing high television viewing (television group 3; 27.6%), and maintaining high television viewing (television group 4; 20.5%). For labeling purposes, we defined low television viewing as less than 2 h/d and high television viewing as 2 h/d or greater. The sex distributions were not different by groups (P = .37): the proportion of females was 53.3% in television group 1, 44.6% in television group 2, 53.6% in television group 3, and 48.2% in television group 4.

Based on the dual trajectory analysis of MVPA and organized sports participation, no participants in sport group 1 (no sport participation) were categorized as consistently active (Table 2). Conversely, no participants in sport groups 2 or 3 (dropout from sports participation or consistent sports participation, respectively) were categorized as consistently inactive. Dual trajectory analysis of MVPA and television viewing indicated that the healthiest television viewing pattern (television group 1, decreasing low television viewing) was associated with the consistently active pattern (P < .001) (Table 2). However, the unhealthiest television viewing pattern (television group 4: maintaining high television viewing) was not associated with the consistently inactive pattern: the proportion of children in MVPA group 1 was not significantly different between television groups 1 and 4 (P = .35) or between television group 4 and the other 3 television groups (P = .27).

Discussion

This study mapped and associated trajectories of 3 ubiquitous behaviors from childhood to young adulthood: MVPA, organized sports participation, and television viewing. It confirmed that most children follow a pattern of decline in moderate MVPA from childhood to young adulthood. Most important, a sizeable number of children, especially girls, are inactive as children and stay that way through adolescence. The MVPA trajectory analysis also provided a new insight via the identification of a subset of participants who are consistently active from childhood into young adulthood. Almost 1 in 5 IBDS participants followed this pattern, and many of them routinely watched little television and, as they aged, watched even less television. The consistently active pattern was not previously identified in an IBDS sex-specific analysis,33 presumably owing to reduced statistical power. The level of MVPA among the consistently active group (approximately 45 min/d) was not as high as the US Physical Activity Guidelines recommend.2 However, the guidelines are primarily based on epidemiologic studies that linked health outcomes data to self-reported PA or time reported during an exercise intervention,2,34 for which both approaches are likely to have a higher count of minutes than are accelerometer measures.35 Therefore, the consistently active group could be the subpopulation that sustains a desirable level of PA. In future studies, an investigation of group-based trajectories of concurrently measured accelerometry PA and self-reported PA will help to translate the US Physical Activity Guidelines and develop guidelines based on device-measured movement.

This study suggests that, although consistent sports participation is insufficient to maintain healthy MVPA behavior, joining organized youth sports (despite eventually dropping out) could be important to avoid a consistently inactive lifestyle. In addition to the potential causal pathway between no organized sports participation and a low MVPA pattern, shared participant characteristics, such as low socioeconomic status,36,37 could partly explain the observed association. In the IBDS population, sex was not a predictor of organized sports participation trajectories. The even distribution by sex in the sports participation trajectory classes is not surprising considering that high schools in Iowa offer girls substantial access to organized sports (high school sports teams accommodate 72% of girls).38 More research is needed to characterize children in the consistently low MVPA and no sports participation trajectories.

The association between healthy MVPA and television viewing behaviors supports the idea that the developmental pathways of healthy MVPA and television viewing behaviors could be related. Previous studies have shown a cluster pattern for PA and sedentary behaviors.39 Some of the studies also found an unhealthy clustering of PA and sedentary behavior.39 In the current study, however, an unhealthy clustering of MVPA and television viewing trajectories was less clear. The development of television viewing behavior among inactive children should be further investigated in future studies.

Several limitations should be acknowledged. First, changes in the PAQ versions might have resulted in a differential bias in estimating organized sports participation and television viewing hours. However, additional group-based trajectory analyses of television viewing and sports participation, including data collected at IBDS waves 3 to 7, produced the same number of distinct subgroups and similar trajectory patterns. Second, because a group-based trajectory model assumes that missing data are random, nonrandom missing data across time could have biased the study findings. However, of those who were included in the data analysis, many (60.3%) completed at least 6 MVPA waves. To our knowledge, the IBDS offers the most complete longitudinal youth accelerometry-measured PA data, including multiple follow-ups that are desirable for group-based trajectory analysis. Third, this study used a sample that was homogeneously white and of high socioeconomic level. Therefore, the results may not be generalizable to other populations. Last, it is important to note that, because group membership was determined based on the maximum likelihood, not all group members perfectly followed their group’s trajectory. Therefore, use of the posterior probability-based classification that does not account for the uncertainty in group classification is prone to classification error of the group40 and should be interpreted with caution. However, we used this method because the posterior probability-based classification provides a straightforward basis for producing profiles of trajectory members and an easy and transparent communication of the findings.22

Despite these limitations, this study has important strengths. This study used the most comprehensive longitudinal data set available for PA, sports participation, and television viewing measurements among youth: MVPA was assessed 7 times and sports participation and television viewing were assessed on average 23 times during 14 years. This study applied an innovative analytic approach—group-based trajectory analysis—that allows for the identification of diverse developmental trajectories of MVPA, sports participation, and television viewing behaviors.

Conclusions

This study provided a nuanced look at the known decrease in MVPA from childhood to young adulthood. We identified a consistently active trajectory pattern that represents a seemingly healthy level of MVPA during childhood to young adulthood. This study suggests that sports participation could be critical to avoid a consistently inactive pattern and that the developmental pathways of PA and television viewing behaviors could be related.

Back to top
Article Information

Accepted for Publication: February 5, 2015.

Corresponding Author: Soyang Kwon, PhD, Stanley Manne Children’s Research Institute, Ann & Robert H. Lurie Children’s Hospital of Chicago, 225 E Chicago Ave, PO Box 157, Chicago, IL 60611 (skwon@luriechildrens.org).

Published Online: May 18, 2015. doi:10.1001/jamapediatrics.2015.0327.

Author Contributions: Dr Kwon had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Kwon, Janz.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Kwon, Janz.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Kwon, Janz, Burns.

Obtained funding: Kwon, Janz, Levy.

Administrative, technical, or material support: Kwon, Janz, Levy.

Study supervision: Kwon.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by grants R03 HD078966, R01 DE12101, R01 DE09551, UL1 RR024979, UL1 TR000442, and M01 RR00059 from the National Institutes of Health, the Roy J. Carver Charitable Trust, and Delta Dental of Iowa Foundation.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank the children, parents, and staff of the Iowa Fluoride Study and the Iowa Bone Development Study. The families who participated in the studies were compensated for their participation.

References
1.
Eime  RM, Young  JA, Harvey  JT, Charity  MJ, Payne  WR.  A systematic review of the psychological and social benefits of participation in sport for children and adolescents: informing development of a conceptual model of health through sport. Int J Behav Nutr Phys Act. 2013;10:98.
PubMedArticle
2.
Physical Activity Guidelines Advisory Committee. Physical Activity Guidelines Advisory Committee Report. Washington, DC: US Dept of Health and Human Services; 2008.
3.
Floriani  V, Kennedy  C.  Promotion of physical activity in children. Curr Opin Pediatr. 2008;20(1):90-95.
PubMedArticle
4.
Troiano  RP, Berrigan  D, Dodd  KW, Mâsse  LC, Tilert  T, McDowell  M.  Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181-188.
PubMedArticle
5.
US Department of Health and Human Services. Promoting Better Health for Young People Through Physical Activity and Sports. Atlanta, GA: The Centers for Disease Control and Prevention; 2000.
6.
Geidne  S, Quennerstedt  M, Eriksson  C.  The youth sports club as a health-promoting setting: an integrative review of research. Scand J Public Health. 2013;41(3):269-283.
PubMedArticle
7.
Spinks  A, Macpherson  A, Bain  C, McClure  R.  Determinants of sufficient daily activity in Australian primary school children. J Paediatr Child Health. 2006;42(11):674-679.
PubMedArticle
8.
Nilsson  A, Andersen  LB, Ommundsen  Y,  et al.  Correlates of objectively assessed physical activity and sedentary time in children: a cross-sectional study (The European Youth Heart Study). BMC Public Health. 2009;9:322.
PubMedArticle
9.
Wickel  EE, Eisenmann  JC.  Contribution of youth sport to total daily physical activity among 6- to 12-yr-old boys. Med Sci Sports Exerc. 2007;39(9):1493-1500.
PubMedArticle
10.
Katzmarzyk  PT, Malina  RM.  Contribution of organized sports participation to estimated daily energy expenditure in youth. Pediatr Exerc Sci. 1998;10(4):378-386.
11.
Sallis  JF, Prochaska  JJ, Taylor  WC.  A review of correlates of physical activity of children and adolescents. Med Sci Sports Exerc. 2000;32(5):963-975.
PubMedArticle
12.
Nelson  TF, Stovitz  SD, Thomas  M, LaVoi  NM, Bauer  KW, Neumark-Sztainer  D.  Do youth sports prevent pediatric obesity? a systematic review and commentary. Curr Sports Med Rep. 2011;10(6):360-370.
PubMedArticle
13.
Sabo  D, Veliz  P. Go Out and Play: Youth Sports in America. East Meadow, NY: Women's Sports Foundation; 2008.
14.
Robinson  TN.  Television viewing and childhood obesity. Pediatr Clin North Am. 2001;48(4):1017-1025.
PubMedArticle
15.
Gebremariam  MK, Bergh  IH, Andersen  LF,  et al.  Are screen-based sedentary behaviors longitudinally associated with dietary behaviors and leisure-time physical activity in the transition into adolescence? Int J Behav Nutr Phys Act. 2013;10:9.
PubMedArticle
16.
Motl  RW, McAuley  E, Birnbaum  AS, Lytle  LA.  Naturally occurring changes in time spent watching television are inversely related to frequency of physical activity during early adolescence. J Adolesc. 2006;29(1):19-32.
PubMedArticle
17.
Koezuka  N, Koo  M, Allison  KR,  et al.  The relationship between sedentary activities and physical inactivity among adolescents: results from the Canadian Community Health Survey. J Adolesc Health. 2006;39(4):515-522.
PubMedArticle
18.
Epstein  LH, Roemmich  JN, Paluch  RA, Raynor  HA.  Physical activity as a substitute for sedentary behavior in youth. Ann Behav Med. 2005;29(3):200-209.
PubMedArticle
19.
Robinson  TN, Hammer  LD, Killen  JD,  et al.  Does television viewing increase obesity and reduce physical activity? cross-sectional and longitudinal analyses among adolescent girls. Pediatrics. 1993;91(2):273-280.
PubMed
20.
Taveras  EM, Field  AE, Berkey  CS,  et al.  Longitudinal relationship between television viewing and leisure-time physical activity during adolescence. Pediatrics. 2007;119(2):e314-e319.
PubMedArticle
21.
Melkevik  O, Torsheim  T, Iannotti  RJ, Wold  B.  Is spending time in screen-based sedentary behaviors associated with less physical activity: a cross national investigation. Int J Behav Nutr Phys Act. 2010;7:46.
PubMedArticle
22.
Nagin  D. Group-Based Modeling of Development. Cambridge, MA: Harvard University Press; 2005.
23.
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):70-77.
PubMedArticle
24.
Janz  KF, Burns  TL, Levy  SM.  Tracking of activity and sedentary behaviors in childhood: the Iowa Bone Development Study. Am J Prev Med. 2005;29(3):171-178.
PubMedArticle
25.
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.
PubMedArticle
26.
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-S111.
PubMed
27.
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):1557-1565.
PubMedArticle
28.
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):1360-1368.
PubMedArticle
29.
Kowlaski  KC, Crocker  PRE, Faulkner  RA.  Validation of the Physical Activity Questionnaire for older children. Pediatr Exerc Sci. 1997;9(2):174-186.
30.
Kowlaski  KC, Crocker  PRE, Kowlaski  NP.  Convergent validity of the Physical Activity Questionnaire for adolescents. Pediatr Exerc Sci. 1997;9(4):342-352.
31.
Jones  B, Nagin  D. A Stata plugin for estimating group-based trajectory models.2012. http://ssrc.indiana.edu/seminars/wimdocs/2013-03-29_nagin_trajectory_stata-plugin-info.pdf. Accessed March 26, 2015.
32.
Nylund  KL, Asparouhov  T, Muthen  BO.  Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Modeling. 2007;14(4):535-569. doi:10.1080/10705510701575396.Article
33.
Janz  KF, Letuchy  EM, Burns  TL, Eichenberger Gilmore  JM, Torner  JC, Levy  SM.  Objectively measured physical activity trajectories predict adolescent bone strength: Iowa Bone Development Study. Br J Sports Med. 2014;48(13):1032-1036.
PubMedArticle
34.
Troiano  RP, McClain  JJ, Brychta  RJ, Chen  KY.  Evolution of accelerometer methods for physical activity research. Br J Sports Med. 2014;48(13):1019-1023.
PubMedArticle
35.
Slootmaker  SM, Schuit  AJ, Chinapaw  MJ, Seidell  JC, van Mechelen  W.  Disagreement in physical activity assessed by accelerometer and self-report in subgroups of age, gender, education and weight status. Int J Behav Nutr Phys Act. 2009;6:17.
PubMedArticle
36.
Vella  SA, Cliff  DP, Okely  AD.  Socio-ecological predictors of participation and dropout in organised sports during childhood. Int J Behav Nutr Phys Act. 2014;11:62.
PubMedArticle
37.
Walters  S, Barr-Anderson  DJ, Wall  M, Neumark-Sztainer  D.  Does participation in organized sports predict future physical activity for adolescents from diverse economic backgrounds? J Adolesc Health. 2009;44(3):268-274.
PubMedArticle
38.
Sabo  D, Veliz  P. Progress Without Equity: The Provision of High School Athletic Opportunity in the United States, by Gender 1993-94 Through 2005-2006. New York, NY: Women's Sports Foundation; 2011.
39.
Leech  RM, McNaughton  SA, Timperio  A.  The clustering of diet, physical activity and sedentary behavior in children and adolescents: a review. Int J Behav Nutr Phys Act. 2014;11:4.
PubMedArticle
40.
Roeder  K, Lynch  K, Nagin  D.  Modeling uncertainty in latent class membership: a case study in criminology. J Am Stat Assoc. 1999;94(447):766-776.Article
×