Background
Routine physical activities occur within ecological contexts. Research must be cognizant of the setting of physical activities and the attributes of physical movements occurring in each context. Most children attend school and spend considerable time in school buildings and on school campuses. Prior research suggests school characteristics may influence physical activity. However, few studies have examined associations between school buildings and campus characteristics and objective measures of physical activity in middle school students.
Objectives
-
Describe the development of objective measures of "active" schools using GIS data, site visits, archival records and aerial photographs (orthophotos).
-
Discuss the development of statistical models defining accelerometer measures of school day physical activity.
-
Present findings regarding associations between objectively measured school characteristics and accelerometer estimates of physical activity in middle school youth.
Methods
The study included 248 participants (42% female, mean age = 13.7) with objectively monitored physical activity data in 1997 were a stratified random sample from 1,295 ethnically diverse seventh and eighth grade students in 10 schools participating in the Planet Health Study. Fifty-six percent of the students were white, 11% black, 14% Hispanic, 11% Asian and 8% other race/ethnicity. Student activity levels were measured using a TriTrac-R3D that was worn for one or two four-day sessions.
School Characteristics
Objective physical characteristics including school campus area, campus play area, building area and area densities per student were assessed retrospectively using local land use parcel data, site visits, orthophotos, architectural plans and site maps.
Building footprints and parcel boundaries were available from local government offices for eight schools. ArcGIS 9 (ESRI) was used to screen-digitize school campus parcel boundaries and building footprints for the remaining two schools based on 2001 orthophotos. Orthophotos as 1:5000 ortho-images from 2001 and 1995 were downloaded from MassGIS, Massachusetts Executive Office of Environmental Affairs. Black and white orthophotos from 1995 defined campus areas where, at two schools, school buildings had been altered between 1997 and 2001 and verified 2001 imagery.
ArcGIS was used to calculate the total area of all campus polygons and to calculate four additional campus area categories for each school (i.e., building footprint, play area, pavement and "other") creating a school site map.
Consultation with school operations staff during site visits provided reported school building area and in conjunction with architectural drawings and site maps, verified site and building characteristics and changes occurring since 1997. Buildings were characterized and subdivided according by floors using architectural plans. Number of building floors and building footprint area calculations were used to estimate GIS-school building areas.
Physical Activity Data
Estimated hierarchical regression models using SAS PROC MIXED were calculated to examine associations between school environmental variables and objective physical activity levels during the school day. The outcome variable, TriTrac-R3D vector magnitude (a measure of movement) was averaged for each 15-minute interval (N=16,619) during the school day and (natural) log-transformed. The model statement included indicator variables for time-of-day (15-minute interval), day-of-week, sex and race/ethnicity. Student age in years, days/week of physical education (PE), and student BMI were included as continuous model covariates. Time-of-day was used as a continuous variable in the “REPEATED” statement. The model “TYPE” statement specified a power spatial covariance matrix.
Results
Total school campus areas ranged from 3,263-129,936 m2 (median=15,989 m2). Campus area/student ranged from 8.8-143.7 m2. GIS-school building area ranged from 5,727-20,312 m2 (median=9,470 m2) and building area/student from 12.1-24.7 m2. Correlation between GIS-calculated and school-reported school building areas was r=.94 (p<.0001). Total play area ranged from 352-48,532 m2 (median=4,941m2) and play area/student from 0.4-58.9m2. Play area comprised from 2.7-62.3% of total campus area across schools. In separate regression models adjusting for student age, sex, race/ethnicity, BMI, PE days/week, day-of-week and time-of-day, the variables school campus area (p<.0001), building area (p=.02), and play area (p=.0012) were each directly associated with log-TriTrac-R3D vector magnitude. An approximate increase of between 20 and 30% in average vector magnitude was associated with the difference in total campus, school and play areas seen in this sample of schools, independent of the other variables in the model. Similar findings were observed in separate models using per student total campus, play and building areas. Compared to schools with fewer than 3.7m2/student of play area, attending a school with greater than 29.5m2/student in play area was associated with an approximate 20% increase in average vector magnitude, (p<0.05). These increases translate into approximately 34 Kcals/day, or walking two extra miles weekly.
Conclusions
Larger school campuses, school buildings and play areas and areas per enrolled student are associated with increased physical activity in middle school students.
Support
This study was supported by funding provided by the Robert Wood Johnson Foundation’s Active Living Research Grant #050376. During the time of study, Joseph Allen was at the Harvard School of Public Health.
