Open Collections

UBC Faculty Research and Publications

Within-person associations of young adolescents’ physical activity across five primary locations: is… Carlson, Jordan A; Mitchell, Tarrah B; Saelens, Brian E; Staggs, Vincent S; Kerr, Jacqueline; Frank, Lawrence D; Schipperijn, Jasper; Conway, Terry L; Glanz, Karen; Chapman, Jim E; Cain, Kelli L; Sallis, James F Apr 20, 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
52383-12966_2017_Article_507.pdf [ 550.17kB ]
Metadata
JSON: 52383-1.0347219.json
JSON-LD: 52383-1.0347219-ld.json
RDF/XML (Pretty): 52383-1.0347219-rdf.xml
RDF/JSON: 52383-1.0347219-rdf.json
Turtle: 52383-1.0347219-turtle.txt
N-Triples: 52383-1.0347219-rdf-ntriples.txt
Original Record: 52383-1.0347219-source.json
Full Text
52383-1.0347219-fulltext.txt
Citation
52383-1.0347219.ris

Full Text

RESEARCH Open AccessWithin-person associations of youngadolescents’ physical activity across fiveprimary locations: is there evidence ofcross-location compensation?Jordan A. Carlson1,2*, Tarrah B. Mitchell3, Brian E. Saelens4, Vincent S. Staggs1,2, Jacqueline Kerr5,Lawrence D. Frank6, Jasper Schipperijn7, Terry L. Conway5, Karen Glanz8, Jim E. Chapman9, Kelli L. Cain5 andJames F. Sallis5AbstractBackground: Youth are active in multiple locations, but it is unknown whether more physical activity in one locationis associated with less in other locations. This cross-sectional study examines whether on days with more physicalactivity in a given location, relative to their typical activity in that location, youth had less activity in other locations(i.e., within-person associations/compensation).Methods: Participants were 528 adolescents, ages 12 to 16 (M = 14.12, SD = 1.44, 50% boys, 70% White non-Hispanic).Accelerometer and Global Positioning System devices were used to measure the proportion of time spent inmoderate-to-vigorous physical activity (MVPA) in five locations: home, home neighborhood, school, schoolneighborhood, and other locations. Mixed-effects regression was used to examine within-person associations ofMVPA across locations and moderators of these associations.Results: Two of ten within-participant associations tested indicated small amounts of compensation, and oneassociation indicated generalization across locations. Higher at-school MVPA (relative to the participant’s average)was related to less at-home MVPA and other-location MVPA (Bs = −0.06 min/day). Higher home-neighborhoodMVPA (relative to the participant’s average) was related to more at-home MVPA (B = 0.07 min/day). Some modelsshowed that compensation was more likely (or generalization less likely) in boys and non-whites or Hispanicyouth.Conclusions: Consistent evidence of compensation across locations was not observed. A small amount ofcompensation was observed for school physical activity, suggesting that adolescents partially compensated forhigh amounts of school activity by being less active in other locations. Conversely, home-neighborhood physicalactivity appeared to carry over into the home, indicating a generalization effect. Overall these findings suggestthat increasing physical activity in one location is unlikely to result in meaningful decreases in other locations.Supporting physical activity across multiple locations is critical to increasing overall physical activity in youth.Keywords: Built environment, Global Positioning Systems (GPS), Neighborhood, School* Correspondence: jacarlson@cmh.edu1Center for Children’s Healthy Lifestyles and Nutrition, Children’s MercyHospital, 610 E. 22nd St., Kansas City, MO 64108, USA2University of Missouri Kansas City, Kansas City, MO, USAFull list of author information is available at the end of the article© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Carlson et al. International Journal of Behavioral Nutritionand Physical Activity  (2017) 14:50 DOI 10.1186/s12966-017-0507-xBackgroundDespite the numerous benefits of physical activity inyouth, current estimates indicate that most adoles-cents are not engaging in adequate amounts of phys-ical activity. Data from 105 countries indicated thatonly 19% of adolescents worldwide achieved the rec-ommended 60 min of physical activity per day basedon self-report [1, 2], and there is evidence that thistrend continues into adulthood [3]. Therefore, moreresearch is needed to better understand patterns ofphysical activity in youth to inform efforts to promotephysical activity.Youth have the potential to be active in multiplelocations, including within schools, homes, neighbor-hoods, and recreation areas such as parks, communitycenters, and sports facilities [4–9]. Because strategiesfor supporting physical activity differ by location,many public health intervention recommendations arelocation-specific (e.g., school-based physical activity,home-based screen time, neighborhood walking) [10–13].Given the goal of increasing overall physical activity, it isimportant to understand whether increased physical activityin one location is related to decreased physical activity inother locations.The compensation hypothesis [14] posits that youthmaintain a presumed activity set-point by compensat-ing for higher than usual activity at one time-point byengaging in lower than usual activity at a later time-point (i.e., compensation). Although several studies inyouth did not find compensation to occur [15–17],there is some evidence suggesting that partial com-pensation occurs within and across days [18, 19].There has been little to no examination of whethercompensation occurs across locations, but this ques-tion is useful to investigate because it has implica-tions for settings-based interventions. Such evidencewould inform intervention strategies and prioritiesand help gauge the potential impact of setting-basedinterventions on overall physical activity. For example,if youth receive more physical activity at school butreduce their activity in other locations, coordinatedmulti-setting strategies may be needed to preventsuch compensation from occurring.The present study investigated how youth’s physicalactivity in one location was associated with physicalactivity in other locations. Specifically, the study examinedwhether an individual youth engaged in more or lessactivity in other locations when he/she was moreactive in a given location relative to his/her typicalactivity in that location (i.e., within-person associa-tions). Moderators of compensation were also investi-gated to determine whether associations differed byparticipant characteristics (e.g., gender, age, neighbor-hood factors).MethodsParticipants and proceduresData were from the Teen Environment and Neighbor-hood (TEAN) study of built environments and physicalactivity conducted in two US regions (Baltimore, MD/Washington, DC and Seattle/King County, WA) during2009–2011. TEAN participants were 928 healthy adoles-cents ages 12–16 years, and one of their parents,selected from 447 census block groups representing highor low walkability and high or low income [20]. Datacollection occurred during the school year and was bal-anced by season across the block group types. Overallparticipation rate (i.e., returned surveys divided byeligible contacts) was 36% and did not vary by neighbor-hood walkability or income. Comparisons of partici-pants’ household demographics with census dataindicated the study sample had higher education andhousehold income compared to residents of the 447 cen-sus block groups in which participants lived. Regardingrace/ethnicity, the study sample was comparable to cen-sus data for adolescent participants, with 34% beingnon-White or Hispanic versus 37% of adolescents in thecensus block groups from which participants were re-cruited. This study was approved by the sponsoringinstitution’s human subjects’ protection committee, par-ents provided informed consent, and adolescents pro-vided assent.Participants were asked to wear an accelerometer andGlobal Positioning System (GPS) tracker during all wak-ing hours for 7 days, except during water activities.Present analyses included a subsample of 528 TEANparticipants. Participants who were not given a GPSdevice (N = 130) and those who did not wear the pro-vided accelerometer and GPS tracker together for ≥1valid school day and ≥1 valid non-school day (N = 148)were excluded. Participants who attended homeschool,did not provide their school address, or had geocodingerrors were excluded (N = 122). Participant demographiccharacteristics and MVPA did not differ significantlybetween the present subsample and the full sample.MeasuresDemographic characteristicsAdolescents’ age, gender, and race/ethnicity (white non-Hispanic vs. non-white or Hispanic) were self-reported,and parents reported the highest level of education (col-lege degree or higher vs. other) attained by any adult inthe household.Accelerometer-measured MVPAAdolescents wore an Actigraph accelerometer on a beltat their left iliac crest during waking hours, with acceler-ation recorded at 30-s epochs. Multiple Actigraphmodels were used (7164, 85.2%; 71256, 5.1%; GT1M,Carlson et al. International Journal of Behavioral Nutrition and Physical Activity  (2017) 14:50 Page 2 of 97.2%; GT3X, 2.5%), however, model type was not associ-ated with MVPA in this study. MVPA was scored usingthe Evenson cut point for youth applied to the verticalaxis acceleration counts [21], which has been shown tohave excellent classification accuracy [22]. Groups of>60 sequential 30-s epochs (i.e., 30 min) with count = 0were considered non-wear [23], and non-wear time wasexcluded from the data. Only days with ≥8 h of validwear time were included.GPS-derived variablesParticipants wore a GlobalSat DG-100 GPS tracker, withlatitude and longitude collected every 30 s when a GPSsignal was attainable. Previous studies documented ac-ceptable performance for tracking participants’ time andlocation patterns in epidemiological studies [24]. ThePersonal Activity and Location Measurement System(PALMS) [25] was used to merge GPS and accelerom-eter data and filter invalid GPS fixes caused by satelliteinterference; the devices were time-synchronized duringinitialization and linked in PALMS using their timestamp. Only days with ≥8 h of GPS signal during accel-erometer wear time were included.Home and school addresses were geocoded and incor-porated into ArcGIS (ESRI, Inc; Redlands, CA) to createspatial buffers, and spatial analyses were performed inPostgreSQL to identify each participant’s amount of timeand MVPA in 5 locations of interest. The locations weredefined as follows: (1) home (50-m radius circular bufferaround the point resulting from geocoding the home ad-dress), (2) home neighborhood (1-km street-network buf-fer around geocoded home address point, excluding theat-home circular buffer), (3) school (15-m buffer aroundgeocoded school parcel), (4) school neighborhood (1-kmstreet-network buffer around geocoded school point,excluding the at-school parcel buffer), and (5) all “other”locations (i.e., any location not included in the aforemen-tioned 4 locations). Participants whose GPS indicated theynever left their home over the monitoring period wereconsidered to have not worn the device and were ex-cluded. Participants who had overlap in their home neigh-borhood and school neighborhood buffers (20% ofsample) were omitted from the analysis comparing home-neighborhood and school-neighborhoods MVPA. For allother models, overlapping time and MVPA were splitevenly across the two overlapping buffers.The resulting variables were computed at the day-leveland were minutes per day of time present and MVPAoccurring in each location. On a given day, if the partici-pant spent 0 min in a location, MVPA for that locationwas scored as 0 min, with the exception of never leavinghome during the monitoring period described above. Be-cause school days and non-school days are distinct withregards to daily activities and patterns, a variable wasderived to denote whether a day was a school or non-school day, with school day defined as any weekday theGPS showed the participant to be within the school par-cel for ≥200 min.AnalysesDescriptive statistics were used to present MVPA andtime across locations, separately for school days and non-school days. Next, day-level regression analyses wereconducted using mixed effects regression to accountfor the nested data structure. Within-person associa-tions were investigated between (a) MVPA in eachlocation (independent variable) and overall MVPA(dependent variable), and (b) MVPA in each possiblepair of locations. The five locations resulted in tencross-location comparisons (each location was com-pared to the other four locations). Each participant’slocation-specific MVPA values for school and non-school days were mean-centered on the participant’saverage MVPA (across days) for that location on thegiven type of day (school or non-school) during theassessment period. This statistical approach forwithin-person associations is similar to creating differ-ences scores (e.g., the participant’s MVPA in locationA on a given day minus the participant’s averageMVPA in location A across days), with the conceptualresearch question being “do days with above averageMVPA in location A have below (i.e., compensation)or above (i.e., what we refer to as generalization)average MVPA in location B?” separately on schoolversus non-school days. All analyses were adjusted forparticipant mean-centered total time in each location.Participant characteristics were not entered as covariatesbecause between-person main effects were eliminated inmean-centering. The models tested within-person effectsand any moderation of these effects by participant charac-teristics. Specifically, participant age, gender, race/ethni-city, and BMI, highest level of parent education, andneighborhood median income and high or low walkability[20] were tested as moderators of the within-person asso-ciations between MVPA in each pair of locations usinginteraction terms. Initially, each model also included aterm to test the interaction between minutes of MVPA inthe location (i.e., the independent variable) and school day(y/n) to explore differences across school days and non-school days. No school day interactions had a p value < .1,so the models reported combined school and non-schooldays. Unstandardized coefficients (B) for MVPA variablesare presented and can be interpreted as the increase or de-crease in the dependent variable (in minutes per day)associated with a 1-min increase in the independent vari-able. Interaction coefficients estimate the change in theMVPA effect (slope) associated with a 1-unit increase inthe moderator.Carlson et al. International Journal of Behavioral Nutrition and Physical Activity  (2017) 14:50 Page 3 of 9ResultsParticipants (N = 528) were from 317 block groups and244 schools. Participants had a mean age of 14.12 (SD =1.44), 50% were girls, 70% were White non-Hispanic,46% lived in a high-walkability neighborhoods, 50% inhigh-income neighborhoods, and 52% resided in theSeattle/King County, WA region. Participants had amean of 42.2 (SD = 22.5) minutes per day of overallMVPA across locations on school days, and 32.3 minper day on non-school days. Participants wore the accel-erometer and GPS devices together for a mean of 3.9(SD = 1.5) valid school days and 3.3 (SD = 1.7) valid non-school days, for a sample total of 3776 days. Location-specific MVPA is presented in Table 1.For each location, days when adolescents had moreMVPA in the location as compared to his/her average inthat location had more overall MVPA (i.e., a 1-min/dayincrease in location-specific MVPA was related to a 0.88–1.03 min/day increase in overall MVPA; Table 2). Aregression coefficient <1.0 indicated that some compen-sation had occurred, whereas a regression coefficient>1.0 indicated that generalization across locations hadoccurred. The largest compensation effect was observedfor at-school MVPA (B = 0.88; i.e., 12% compensation).The coefficients for home-neighborhood and school-neighborhood MVPA indicated that each minute of home-neighborhood MVPA was associated with 1.03 min/dayof overall MVPA, and each minute/day of school-neighborhood MVPA was associated with 1.01 min/dayof overall MVPA (i.e., 1–3% generalization).Two of the ten cross-location comparisons indicatedcompensation, and one of the 10 indicated generalization(Table 3). On days when adolescents had more at-schoolMVPA relative to their average at-school MVPA, they hadless at-home MVPA and other-location MVPA (both Bs= -0.06 min/day, i.e., each 6% compensation). On dayswhen adolescents had more home-neighborhood MVPArelative to their average, they had more at-home MVPA(B = 0.07 min/day; i.e., 7% generalization).Findings regarding whether participant factors moder-ated associations of MVPA between each pair oflocations are summarized in Table 3 and presented infull detail in the Additional file 1: Table S1. Gender wasa significant moderator in three of the 10 models, withgirls showing generalization and boys no effect in twomodels, and boys showing compensation and girls no ef-fect in one model (see Fig. 1). Race/ethnicity was a sig-nificant moderator in two models, with non-whites orHispanics compensating more than white non-Hispanics.Neighborhood income, child age, and neighborhoodwalkability each emerged as moderators in only oneof the 10 models.DiscussionThe present study found limited consistent evidence ofcompensation or generalization of adolescent physicalactivity across locations. Thus, amounts of physical ac-tivity in multiple locations were mainly independentfrom each other. Small compensation effects were ob-served for physical activity between the school locationand at-home/other locations (12% total), whereas a smallgeneralization effect was observed between the homeneighborhood and at-home location (7%). Results indi-cate the home neighborhood is a promising location forinterventions targeting physical activity in adolescents,without concern of compensation. Programs targetingschool physical activity may not result in a 1:1 contribu-tion to overall physical activity (i.e., each minute ofschool physical activity may not equate to a full minuteTable 1 Young adolescents’ time and MVPA by location (N = 528 participants)Mean (SD) minutes/day of MVPA occurring in location Mean (SD) proportion of overall wear timea spent in locationSchool days Non-school days School days Non-school daysAt home 5.5 (6.7) 12.0 (14.2) 20.3% (15.5) 47.2% (34.8)Home neighborhood 5.5 (9.3) 6.8 (11.6) 9.5% (12.6) 19.6% (28.2)At school 23.2 (15.1) 0.6 (2.3) 57.7% (14.8) 1.2% (2.6)School neighborhood 2.3 (4.3) 1.7 (5.0) 3.1% (5.1) 4.4% (12.8)Other locations 5.6 (9.1) 10.9 (15.3) 9.3% (9.1) 27.3% (26.1)All locations (Overall) 42.2 (22.5) 32.3 (21.8) 100% 100%Note: Means and SDs were calculated across participantsaRefers only to the portion of the day the measurement devices were worn; Mean = 13.3, SD = 1.8 h per dayTable 2 Within-person associations (i.e., compensation) betweenMVPA in each location and overall MVPA (N = 3776 days)Associations with overall MVPAB (SE) minutes/day pAt home MVPA 0.95 (0.03) <.001Home neighborhood MVPA 1.03 (0.04) <.001At school MVPA 0.88 (0.03) <.001School neighborhood MVPA 1.01 (0.09) <.001Other locations MVPA 0.90 (0.03) <.001Note: All models were adjusted for participant age, gender, race/ethnicity, andBMI, parent education, neighborhood income and walkability, time in location,and average (across days of monitoring) time and MVPA in locationCarlson et al. International Journal of Behavioral Nutrition and Physical Activity  (2017) 14:50 Page 4 of 9of overall physical activity because of potential compen-sation in other locations) but still provide meaningfulcontributions.There was no consistent evidence that compensationacross locations was moderated by participant age (al-though the age range was narrow), participant BMI,neighborhood walkability, neighborhood income, orparent education. There was some evidence (3 of 10 in-teractions tested) of moderation by gender, with boysbeing more likely to compensate than girls in one ofthe location comparisons and less likely to showgeneralization in two of the location comparisons.There was some evidence (2 of 10 interactions tested)that adolescents who identified as white non-Hispaniccompensated less (or showed more generalization) thantheir counterparts. Two of the three gender interac-tions, and both of the race/ethnicity interactions in-cluded at-home physical activity. It could be that homephysical activity is less structured and less social thanphysical activity that occurs elsewhere, making it andits propensity for being compensated for more likely tobe influenced by individual-level factors. Althoughamount of physical activity in each location differedacross school and non-school days (presented in moredetail in a previous publication) [9], compensationacross locations did not differ between school and non-school days. This suggests that whether a child attendsschool or not on a given day does not appear to be rele-vant to concerns about compensation. This was some-what surprising given that time in school on schooldays is considerable and could conceivably make com-pensation across non-school locations more likely.Most studies that have investigated compensationhave done so across time, especially days, rather thanacross locations [15–19]. These studies have had mixedresults, but some have found evidence of compensation[16, 18, 19, 26]. One study found that children partiallycompensated within days (across time periods) and be-tween days by engaging in fewer steps directly followinga day (or time period) with more than average steps[19]. Comparability of previous findings on compensa-tion is challenging because unlike the present study,prior studies did not account for natural variations in timeuse by adjusting for the total amount of time participantsspent in specific locations. Present findings provided lim-ited evidence of compensation or generalization, with theonly evidence of compensation effects being specific tothe school location. Thus, these effects are unlikely tohave major impacts on youth’s overall physical activity.This finding suggests that, although each minute ofschool-based physical activity may not translate to a fulladditional minute of overall physical activity, increasingphysical activity opportunities in school is an importantstrategy for improving overall physical activity, which isin agreement with previous studies and national recom-mendations [12, 27, 28]. To address the potential forcompensation, parents should encourage and supporttheir child to be physically active outside of school onall days, including days with Physical Education (PE) orschool sports.Table 3 Within-person associations (i.e., compensation) among MVPA minutes/day across 5 primary locations (N = 3776 days)Compensation effect Factors p < .05 associatedwith more compensation orless generalization (interaction B)aB (95% CI) minutes/day pAssociations across locationsHome neighborhood→ At home 0.07 (0.04, 0.10) <.001 High income (-0.09)Non-White or Hispanic (-0.10)+1 year in age (-0.03)Home neighborhood→ School neighborhoodb 0.00 (−0.01, 0.02) .806 [none]Home neighborhood→Other locations 0.01 (−0.04, 0.06) .627 Boys (−0.14)At school→ At home −0.06 (−0.09, −0.03) <.001 Boys (−0.07)Non-White or Hispanic (-0.11)At school→ Home neighborhood −0.02 (−0.05, 0.01) .170 [none]At school→ School neighborhood 0.01 (−0.01, 0.02) .483 [none]At school→Other locations −0.06 (−0.11, −0.02) .006 [none]At home→ School neighborhood −0.02 (−0.04, 0.00) .106 [none]Other locations→ School neighborhood 0.00 (−0.01, 0.01) .861 [none]Other locations→ At home 0.00 (−0.04, 0.03) .762 Boys (−0.06)Higher walkability (−0.10)Note: The independent variable appears before the arrow and the dependent variable appears after the arrow. Daily MVPA in each location was participant meancentered so that the effects would reflect within person differences. All models were adjusted for daily time in location which was also participant mean centeredaModerators tested were participant gender, age, race/ethnicity, and BMI percentile, neighborhood walkability and income, and parent educationbExcluded participants with overlap between their home and school neighborhood (20% of sample)Carlson et al. International Journal of Behavioral Nutrition and Physical Activity  (2017) 14:50 Page 5 of 9Interestingly, generalization of physical activity acrosslocations occurred between the home neighborhood andat-home location. This association could have been dueto activities that took place in and/or directly outside ofthe home and carried over into the neighborhood (i.e.,crossed over the home buffer into the neighborhoodbuffer), such as playing outdoors and active travel. Giventhat home-neighborhood physical activity showed ageneralization effect and that youth are more likely to bephysically active when in their neighborhood than whenat home, at school, and in other locations [9], supportingneighborhood-based physical activity appears to be aparticularly promising strategy for increasing overallphysical activity in youth. Strategies that increaseneighborhood-based physical activity include outdoorwalking, neighborhood play and “Play Streets,” [29], andactive travel to school. Review papers have shown thatactive travel to school does not lead to compensation inphysical activity [30, 31]. Consistent evidence is accumu-lating on the importance of supporting active travel inyouth [32–35], and Safe Routes to School programs wereeffective in several recent evaluations [36–39]. Morewidespread implementation of Safe Routes to Schoolprograms, as well as other evidence-based efforts, areFig. 1 Gender differences in within-person associations among MVPA minutes/day across locationsCarlson et al. International Journal of Behavioral Nutrition and Physical Activity  (2017) 14:50 Page 6 of 9needed to increase the currently low rates of neighbor-hood physical activity and active travel in youth asshown by the present study [40] and others [30, 40–42].Effective strategies are likely multilevel and includecombinations of built environment, neighborhoodsafety, and social/interpersonal strategies [43]. Thepresent finding that physical activity may generalizeacross locations supports comprehensive systems ap-proaches such as those that target multiple strategies inmultiple locations [44, 45].The compensation effects that were found were specificto school-based activity, which has been observed in pre-vious studies [26], whereas the generalization effect wasspecific to home-neighborhood physical activity. It is pos-sible that compensation is more likely to occur when thephysical activity is organized, structured, or even man-dated (e.g., physical education class) than when physicalactivity is discretionary, because discretionary physicalactivity may be less likely to be perceived as “exercise”. Forexample, neighborhood activity is most likely to consist ofactive play and/or active transportation, activities per-formed for fun or out of necessity to accomplish anotherobjective (e.g., getting to school or other destinations) andless likely to be viewed as “exercise”. School-based activitycould include sports, which are likely perceived as highlyactive. Adolescents may engage in lower activity levelsoutside of sports to retain energy for sports, or be inactiveafter participating in sports because of fatigue. Parents orchildren may think that if the child participated in sportsor PE then he/she does not need to be active at home.These beliefs can be detrimental because, based on thepresent and other findings, adolescents accrue less thanhalf of the recommended 60 min per day of physical activ-ity at school.Strengths, limitations, and research gapsThe present study utilized a large sample of adolescentsin two US regions, within-person analyses, and location-specific estimates of objectively-assessed physical activityderived from GPS and accelerometers, which weremethodological strengths. Going beyond main effects,the present study used interaction tests, which revealedno differences between school days and non-school days,with gender as the only consistent moderator of com-pensation effects across locations. These strengths en-hance confidence in the novel contribution of examiningthe possibility of cross-location compensation. For limi-tations, the use of 30-s epochs with accelerometerscould have underestimated participants’ physical activity,as some authors recommend shorter epochs [22, 46].Another limitation was the potential for misclassificationfrom the GPS, such as when signals were unreliable insome indoor environments, and the observational natureof the study, which limited understanding of causality.Since “other locations” were simply any location outsideof the home and school neighborhoods, future studiescould provide more specificity by documenting physicalactivity in key types/categories of locations, such asrecreation locations. Sports facilities, parks, and commu-nity centers are known to be important locations forphysical activity and should be investigated in futurestudies [47]. Another limitation was that indoor physicalactivity may have been underestimated because we omit-ted time with missing GPS information (e.g., due tosignal loss) from the present analyses. Methods haverecently been validated for imputing missing geocoordi-nate information and should be considered in futurestudies [48].Overall, limited evidence of compensation orgeneralization across locations was found, but futurestudies could investigate whether compensation differsby activity type (e.g., sports vs. other school-basedactivity), which was not feasible in the present study.Future studies should also investigate whether youthcompensate for high amounts of physical activity inone location by having more than typical sedentarytime in other locations, which is plausible and hasimportant health implications. The present study didnot investigate compensation by time (e.g., within oracross days), which has been observed in some studies[18, 19], but findings have been inconsistent [15–17].ConclusionsThe primary finding was that adolescent physical activ-ity in one location was mainly independent of activityin other locations. Thus, promoting physical activity inall locations could be expected to contribute toincreased activity levels. However, small amounts ofboth compensation and generalization across locationswere observed. Compensation was observed betweenschool and at-home/other locations, and boys, andnon-whites or Hispanics tended to be more likely tocompensate or less likely to show generalization acrosslocations. The generalization effect was specific toneighborhood-based physical activity. Overall, thesefindings indicated that increasing physical activity inone location is not likely to result in meaningful de-creases in other locations. Supporting physical activityacross multiple locations is critical to increasing overallphysical activity in youth. Targeting neighborhood-based activity should be a high priority, given its criticalrole in overall levels of physical activity.Additional fileAdditional file 1: Table S1. Interactions of within-person associations(i.e., compensation) among MVPA minutes/day across 5 primary locations(N = 3776 days). (DOC 41 kb)Carlson et al. International Journal of Behavioral Nutrition and Physical Activity  (2017) 14:50 Page 7 of 9AbbreviationsGPS: Global positioning system; MVPA: Moderate-to-vigorous physicalactivity; PALMS: Personal activity and location measurement system;TEAN: Teen environment and neighborhoodAcknowledgementsNone.FundingThis study was funded by National Institutes of Health (NIH) grant HL083454.Availability of data and materialsThe data and research materials can be obtained by contacting the first andlast author of this paper.Authors’ contributionsJAC and TBM conceptualized the aims and drafted the initial manuscript; VSScarried out the statistical analyses; JK, BES, LDF, KG, TLC, and JFS conceptualizedand designed the TEAN study; JEC led some of the GIS data collection andanalyses; GS assisted with the GPS analyses; KLC coordinated and superviseddata collection; and all authors critically reviewed and approved the finalmanuscript as submitted.Competing interestsThe authors declare that they have no competing interests.Consent for publicationThe authors consent to have this paper published in IJBNPA.Ethics approval and consent to participateThis study was approved by the sponsoring institution’s human subjects’protection committee and all participants provided informed consent andassent.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1Center for Children’s Healthy Lifestyles and Nutrition, Children’s MercyHospital, 610 E. 22nd St., Kansas City, MO 64108, USA. 2University of MissouriKansas City, Kansas City, MO, USA. 3University of Kansas, Lawrence, KS, USA.4Seattle Children’s Research Institute and the University of Washington,Seattle, WA, USA. 5University of California San Diego, La Jolla, CA, USA.6University of British Columbia, Vancouver, BC, Canada. 7University ofSouthern Denmark, Odense, Denmark. 8University of Pennsylvania,Philadelphia, PA, USA. 9Urban Design 4 Health, Rochester, NY, USA.Received: 2 December 2016 Accepted: 7 April 2017References1. Hallal PC, Andersen LB, Bull FC, Guthold R, Haskell W, Ekelund U. Globalphysical activity levels: surveillance progress, pitfalls, and prospects. Lancet.2012;380:247–57.2. Sallis JF, Bull F, Guthold R, Heath GW, Inoue S, Kelly P, Oyeyemi AL, PerezLG, Richards J, Hallal PC. Progress in physical activity over the Olympicquadrennium. Lancet. 2016;388:1325–36.3. Telama R. Tracking of physical activity from childhood to adulthood: areview. Obes Facts. 2009;2:187–95.4. Grow HM, Saelens BE, Kerr J, Durant NH, Norman GJ, Sallis JF. Where areyouth active? Roles of proximity, active transport, and built environment.Med Sci Sports Exerc. 2008;40:2071–9.5. Kneeshaw-Price S, Saelens BE, Sallis JF, Glanz K, Frank LD, Kerr J, Hannon PA,Grembowski DE, Chan KC, Cain KL. Children’s objective physical activity bylocation: why the neighborhood matters. Pediatr Exerc Sci. 2013;25:468–86.6. Rainham DG, Bates CJ, Blanchard CM, Dummer TJ, Kirk SF, Shearer CL.Spatial classification of youth physical activity patterns. Am J Prev Med.2012;42:e87–96.7. Jones AP, Coombes EG, Griffin SJ, van Sluijs EM. Environmental supportivenessfor physical activity in English schoolchildren: a study using Global PositioningSystems. Int J Behav Nutr Phys Act. 2009;6:42.8. Klinker CD, Schipperijn J, Christian H, Kerr J, Ersboll AK, Troelsen J. Usingaccelerometers and global positioning system devices to assess gender andage differences in children’s school, transport, leisure and home basedphysical activity. Int J Behav Nutr Phys Act. 2014;11:8.9. Carlson JA, Schipperijn J, Kerr J, Saelens BE, Natarajan L, Frank LD, Glanz K,Conway TL, Chapman JE, Cain KL, Sallis JF. Locations of physical activity asassessed by GPS in young adolescents. Pediatrics. 2016;137:1–10.10. U.S. Department of Health and Human Services. 2008 physical activityguidelines for Americans. Available at: http://www.health.gov/paguidelines/pdf/paguide.pdf. Accessed 3 Feb 2017.11. Healthy People 2020. [ http://www.healthypeople.gov/2020/topicsobjectives2020/ ]. Accessed 3 Feb 2017.12. Institute of Medicine. Educating the student body: taking physical activityand physical education to school. Washington, DC: The National AcademiesPress; 2013.13. Pate RR, Davis MG, Robinson TN, Stone EJ, McKenzie TL, Young JC.Promoting physical activity in children and youth: a leadership role forschools: a scientific statement from the American Heart Association Councilon Nutrition, Physical Activity, and Metabolism (Physical Activity Committee)in collaboration with the Councils on Cardiovascular Disease in the Youngand Cardiovascular Nursing. Circulation. 2006;114:1214–24.14. Rowland TW. The biological basis of physical activity. Med Sci Sports Exerc.1998;30:392–9.15. Goodman A, Mackett RL, Paskins J. Activity compensation and activitysynergy in British 8–13 year olds. Prev Med. 2011;53:293–8.16. Gomersall SR, Rowlands AV, English C, Maher C, Olds TS. The ActivityStathypothesis: the concept, the evidence and the methodologies. Sports Med.2013;43:135–49.17. Baggett CD, Stevens J, Catellier DJ, Evenson KR, McMurray RG, He K, Treuth MS.Compensation or displacement of physical activity in middle-school girls: theTrial of Activity for Adolescent Girls. Int J Obes (Lond). 2010;34:1193–9.18. Ridgers ND, Timperio A, Cerin E, Salmon J. Compensation of physicalactivity and sedentary time in primary school children. Med Sci Sports Exerc.2014;46:1564–9.19. Ridgers ND, Timperio A, Cerin E, Salmon J. Within- and between-dayassociations between children’s sitting and physical activity time. BMCPublic Health. 2015;15:950.20. Frank LD, Sallis JF, Saelens BE, Leary L, Cain K, Conway TL, Hess PM. Thedevelopment of a walkability index: application to the NeighborhoodQuality of Life Study. Br J Sports Med. 2010;44:924–33.21. Evenson KR, Catellier DJ, Gill K, Ondrak KS, McMurray RG. Calibration of twoobjective measures of physical activity for children. J Sports Sci. 2008;26:1557–65.22. Trost SG, Loprinzi PD, Moore R, Pfeiffer KA. Comparison of accelerometercut points for predicting activity intensity in youth. Med Sci Sports Exerc.2011;43:1360–8.23. Cain KL, Sallis JF, Conway TL, Van Dyck D, Calhoon L. Using accelerometersin youth physical activity studies: a review of methods. J Phys Act Health.2013;10:437–50.24. Wu J, Jiang C, Liu Z, Houston D, Jaimes G, McConnell R. Performances ofdifferent global positioning system devices for time-location tracking in airpollution epidemiological studies. Environ Health Insights. 2010;4:93–108.25. Physical Activity Location Measurement System. [ http://cwphs.ucsd.edu/palms ]. Accessed 3 Feb 2017.26. Fremeaux AE, Mallam KM, Metcalf BS, Hosking J, Voss LD, Wilkin TJ. Theimpact of school-time activity on total physical activity: the activitystathypothesis (EarlyBird 46). Int J Obes (Lond). 2011;35:1277–83.27. Centers for Disease Control and Prevention. Comprehensive school physicalactivity programs: a guide for schools. Atlanta: US Department of Healthand Human Services; 2013.28. American Alliance for Health PE, Recreation and Dance. Comprehensiveschool physical activity programs: helping students achieve 60 minutes ofphysical activity each day. Reston: 2013.29. D’Haese S, Van Dyck D, De Bourdeaudhuij I, Deforche B, Cardon G. Organizing“Play Streets” during school vacations can increase physical activity anddecrease sedentary time in children. Int J Behav Nutr Phys Act. 2015;12:14.30. Faulkner GEJ, Buliung RN, Flora PK, Fusco C. Active school transport, physicalactivity levels and body weight of children and youth: a systematic review.Prev Med. 2009;48.Carlson et al. International Journal of Behavioral Nutrition and Physical Activity  (2017) 14:50 Page 8 of 931. Larouche R, Saunders TJ, Faulkner G, Colley R, Tremblay M. Associationsbetween active school transport and physical activity, body composition,and cardiovascular fitness: a systematic review of 68 studies. J Phys ActHealth. 2014;11:206–27.32. Lubans DR, Boreham CA, Kelly P, Foster CE. The relationship between activetravel to school and health-related fitness in children and adolescents: asystematic review. Int J Behav Nutr Phys Act. 2011;8:5.33. Cooper AR, Page AS, Foster LJ, Qahwaji D. Commuting to school: arechildren who walk more physically active? Am J Prev Med. 2003;25:273–6.34. Cooper AR, Andersen LB, Wedderkopp N, Page AS, Froberg K. Physicalactivity levels of children who walk, cycle, or are driven to school. Am J PrevMed. 2005;29:179–84.35. Berrigan D, Troiano RP, McNeel T, DiSogra C, Ballard-Barbash R. Activetransportation increases adherence to activity recommendations. Am J PrevMed. 2006;31:210–6.36. McDonald NC, Steiner RL, Lee C, Rhoulac Smith T, Zhu X, Yang Y. Impact ofthe safe routes to school program on walking and bicycling. J Am PlannAssoc. 2014;80:153–67.37. Muennig PA, Epstein M, Li G, DiMaggio C. The cost-effectiveness of New YorkCity’s safe routes to school program. Am J Public Health. 2014;104:1294–9.38. Stewart O, Moudon AV, Claybrooke C. Multistate evaluation of safe routes toschool programs. Am J Health Promot. 2014;28:S89–96.39. Impact of Safe Routes to School Programs on Walking and Biking. [ http://activelivingresearch.org/SRTSreview ]. Accessed 3 Feb 2017.40. Carlson JA, Saelens BE, Kerr J, Schipperijn J, Conway TL, Frank LD, ChapmanJE, Glanz K, Cain KL, Sallis JF. Association between neighborhood walkabilityand GPS-measured walking, bicycling and vehicle time in adolescents.Health Place. 2015;32:1–7.41. McDonald NC. Active transportation to school: trends among U.S.schoolchildren, 1969–2001. Am J Prev Med. 2007;32:509–16.42. Ham SA, Macera CA, Lindley C. Trends in walking for transportation in theUnited States, 1995 and 2001. Prev Chronic Dis. 2005;2:A14.43. Sallis JF, Owen N, Fisher E. Ecological models of health behavior. In: Glanz K,Rimer BK, Viswanath K, editors. Health behavior: theory, research, andpractice. 5th ed. San Francisco: Jossey-Bass; 2015. p. 43–64.44. President’s Council on Fitness Sports and Nutrition. Physical activityguidelines for Americans mid-course report: strategies to increase physicalactivity among youth. Washington: U.S. Dept. of Health and HumanServices; 2013.45. Institute of Medicine. Accelerating progress in obesity prevention: solving theweight of the nation. Washington, D.C.: The National Academies Press; 2012.46. Baquet G, Stratton G, Van Praagh E, Berthoin S. Improving physical activityassessment in prepubertal children with high-frequency accelerometrymonitoring: a methodological issue. Prev Med. 2007;44:143–7.47. Klinker C, Schipperijn J, Kerr J, Ersbøll A, Troelsen J. Context-specific outdoortime and physical activity among school-children across gender and age:using accelerometers and GPS to advance methods. Front Public Health.2014;2:20.48. Meseck K, Jankowska MM, Schipperijn J, Natarajan L, Godbole S, Carlson J,Takemoto M, Crist K, Kerr J. Is missing geographic positioning system datain accelerometry studies a problem, and is imputation the solution?Geospat Health. 2016;11:403.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Carlson et al. International Journal of Behavioral Nutrition and Physical Activity  (2017) 14:50 Page 9 of 9

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.52383.1-0347219/manifest

Comment

Related Items