A longitudinal analysis of relationships between neighborhood context and underserved children's physical activity in a rapidly growing city

Abstract

Neighborhood context, which may be impacted by urban growth or residential mobility, is associated with childhood physical activity. This secondary analysis examined associations of objectively measured neighborhood characteristics with young children's moderate-to-vigorous physical activity (MVPA) and sedentary/rest time (SRT) over a period of rapid infrastructure change. Underserved preschoolers (n = 426) from a 36-month obesity prevention intervention were included in a secondary analysis (2019-2020). Based on household addresses, participants were coded as movers or non-movers and linked to four neighborhood variables: 1) distance to recreation sites, 2) annual crimes, 3) annual stray dogs, and 4) Gini index of income inequality. Accelerometry captured MVPA and SRT at baseline and 36 months. Baseline-to-follow-up neighborhood variables within moved and non-moved groups were compared. Multivariable regression assessed associations between follow-up MVPA/SRT and neighborhood variables. 45.3% of participants (n = 193) moved. Distance to the closest recreation site decreased significantly for non-movers (0.75 to 0.72 mi, p < 0.001). Nearby crimes significantly decreased for both groups (movers: 90 to 80, p < 0.001; non-movers: 77 to 74, p < 0.001) as did stray dogs (movers: 36 to 15, p < 0.001; non-movers: 36 to 18, p < 0.001). Neighborhood income inequality decreased significantly for movers (0.41 to 0.38, p = 0.03). Child MVPA minutes/day significantly decreased over time from median = 84.7 [Q1 = 64.1, Q3 = 103.9] to median = 73.6 [Q1 = 56.1, Q3 = 96.0], p < 0.001). No significant associations were detected between neighborhood variables and child physical activity. In a rapidly growing county, neighborhood context generally improved over time regardless of move status. Within this context, no associations between neighborhood characteristics and MVPA/SRT were detected in children.

Keywords: Built environment; Community based research; Families; Physical activity; Underserved children.

Fig. 1

Selection of cohort for GIS analysis from full GROW trial sample.

Fig. 2

Adjusted multivariable regression models evaluating associations between neighborhood context and child SRT/MVPA. Models controlled for baseline percent MVPA, follow-up accelerometer wear time (minutes), child age (years), child sex, child BMI (kg/m2), parent education, move status, intervention group, and parent ethnicity. These models failed to detect significant associations between neighborhood context variables and child physical activity. The standardized regression coefficients were nearly zero, with 95% confidence intervals crossing zero.