Background and Purpose
Neighborhood environment characteristics are consistently related to adult’s physical activity; specifically, more active residents live in neighborhoods with built environments that encourage physical activity (1). Characterizing neighborhoods is complex, involving various neighborhood aspects of walkability, recreation, transit, aesthetics, and safety and their numerous combinations. Latent profile analysis (LPA), a relatively new analytical technique, has shown promise in identifying combinations or patterns of environmental aspects related to objectively measured moderate-to-vigorous physical activity (MVPA) in both adults and children (2,3,4). Children attain most of their daily physical activity outside of school, thus it is important to understand how the out-of-school environment relates to this activity. However, researchers have yet to examine how patterns in perceived neighborhood environment characteristics relate to children’s out-of-school MVPA (MVPA-OS).

Objectives
This analysis (a) explored latent profile structures underlying parents’ perceptions of neighborhood characteristics in San Diego County (CA) and King County (Seattle, WA) and (b) examined differences in children’s MVPA-OS across empirically derived neighborhood profiles.

Methods
The baseline wave of the Neighborhood Impact on Kids Study5, provided parents’ self-reported neighborhood characteristics and children’s accelerometer-based MVPA data.  Households in San Diego County (CA) and King County (Seattle, WA) with children aged 6-11 years old were randomly sampled from block groups (neighborhoods) selected to represent supportive and unsupportive physical activity and nutrition environments (for details see Frank et al. (2012)6). Parents (n=730) completed the Neighborhood Environment Walkability Scale (NEWS) and items about household, parent, and child demographics. The final sample included 329 children (9.2±1.6 yrs, 49.5% girls, 17.6% nonwhite) from the San Diego region and 373 children (8.9±1.5, 50.1% girls, 16.6% nonwhite) from the Seattle region. We used LPA, a novel technique that maximizes between-group variance and minimizes within-group variance across sets of continuous indicators based on model fit criteria7. Types of neighborhoods were identified from patterns of z-scored estimated means of indicators of perceived residential density, land-use mix diversity, land-use mix access, street connectivity, pedestrian facilities, aesthetics, traffic safety, crime safety, transit access, and parks and recreation facility access. MVPA-OS, measured by ActiGraph (GT1M) accelerometer, was obtained by subtracting children’s unique school-time MVPA, identified by place and time logs, from total daily MVPA. MVPA and MVPA-OS data were scored using Evenson cut points8 based on valid wear (=10 hours) days. We used random effects regression models to examine differences in mean daily MVPA and MVPA-OS across latent profiles, adjusting for demographics (child gender, race/ethnicity, and age; parent marital status, income and education) number of cars per legal driver in household, time lived at current address and nesting (non-independence) within block groups.

Results
Model fit criteria (i.e, AIC and BIC) for the LPAs supported a 4-profile solution for San Diego (n=329) and a 3-profile solution for Seattle (n=373). Profiles were similar across regions except for the Low Walkable/Safe profile (SD-LWS; n=46, 14.0% of sample) present in San Diego (Figure 1) but not in Seattle (Figure 2). In San Diego, the profiles included: Low Walkable/Unsafe/Recreationally Sparse (SD-LWURS; n=77, 23.4%), Moderately Walkable (SD-MW; n=143, 43.5%), and Overall Activity Friendly (SD-OAF; n=63, 19.1%). In Seattle, the profiles included:  Low Walkable/Transit & Recreationally Sparse (S-LWTRS; n=88, 23.6%), Moderately Walkable (S-MW; n=137, 36.7%), and Overall Activity Friendly (S-OAF; n=148, 39.7%). ANCOVA models revealed significantly lower (mean min/day, 95% CI) San Diego child MVPA-OS in the SD-LWURS profile (58.9, 47.2-73.5) compared to the SD-LWS (75.1, 60.1-93.9), SD-MW (71.8, 59.0-87.5), and SD-OAF (73.9, 60.0-91.1) profiles after adjusting for potential confounders. The differences between Seattle profiles were not significant.

Conclusions
Derived neighborhood profiles in San Diego and Seattle regions based on parents’ perceptions of neighborhood characteristics were generally similar, but the unequal number of profiles and differences in some patterns suggests unique aspects in the perceived environment between these regions. Differences in the San Diego region suggested children living in Low Walkable/Unsafe/Recreation Sparse neighborhoods obtained less MVPA-OS than other neighborhoods, while Seattle-area children’s out-of-school physical activity did not differ based on profile membership. Parent-perceived neighborhood characteristics may play a key role in children’s out-of-school physical activity and identifying patterns of neighborhood characteristics that promote or hinder child physical activity is important for improving neighborhood design.

Implications for Practice and Policy
These results suggest that patterns of neighborhood characteristics play a role in child physical activity outside of the school setting, and the patterns may be region-specific. Out-of-school physical activity needs to be addressed through providing opportunities in neighborhoods that are supportive of all children’s physical activity.

References

1. Saelens BE, Sallis JF, Black JB, Chen D. Neighborhood-based differences in physical activity: an environment scale evaluation. Am J Public Health. 2003;93(9):1552–1558.
2. Norman GJ, Adams MA, Kerr J, Ryan S, Frank LD, Roesch SC. A latent profile analysis of neighborhood recreation environments in relation to adolescent physical activity, sedentary time, and obesity. J Public Heal Manag Pr JPHMP. 2010;16(5):411–419. doi:10.1097/PHH.0b013e3181c60e92.
3. Adams MA, Sallis JF, Kerr J, et al. Neighborhood environment profiles related to physical activity and weight status: a latent profile analysis. Prev Med. 2011;52(5):326–331. doi:10.1016/j.ypmed.2011.02.020.
4. McDonald K, Hearst M, Farbakhsh K, et al. Adolescent physical activity and the built environment: a latent class analysis approach. Health Place. 2012;18(2):191–198. doi:10.1016/j.healthplace.2011.09.004.
5. Saelens BE, Sallis JF, Frank LD, et al. Obesogenic neighborhood environments, child and parent obesity: the Neighborhood Impact on Kids study. Am J Prev Med. 2012;42(5):e57–64. doi:10.1016/j.amepre.2012.02.008.
6. Frank LD, Saelens BE, Chapman J, et al. Objective assessment of obesogenic environments in youth: geographic information system methods and spatial findings from the Neighborhood Impact on Kids study. Am J Prev Med. 2012;42(5):e47–55. doi:10.1016/j.amepre.2012.02.006.
7. Collins LM. Latent class and latent transition analysis: with applications in the social behavioral, and health sciences. Hoboken, N.J: Wiley; 2010.
8. 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. doi:10.1080/02640410802334196.

Support / Funding Source
This work was supported by in part by the American Heart Association’s Beginning Grant in Aid (#12BGIA9280017), the NIH National Institute of Environmental Health Sciences (ES014240), USDA 2007-55215-17924, and by grants to the Seattle Children’s Pediatrics.