Exploring the Link between Features of the Neighbourhood Built Environment and Cardiorespiratory Fitness in Youth: An Ensemble Learning Approach Using Random Forests
Cardiorespiratory fitness (CRF) in youth is associated with cardiometabolic outcomes and risk of early-onset cardiovascular disease. While the role of the neighbourhood built environment (NBE) in supporting physical activity is recognized, the contribution of the NBE to CRF requires further investigation. We examined the extent to which a range of NBE features predict youth CRF, using oxygen consumption per unit of fat-free mass as an indicator of the oxidative capacity of lean tissue. Data were from the baseline visit of the Quebec Adipose and Lifestyle Investigation in Youth (QUALITY) cohort study (n=504, ages 8–10 years, Greater Montreal Area, Canada). CRF was defined as peak oxygen consumption during a cycling test on an electromagnetic bike, adjusted for Fat-Free Mass (VO2peak/FFM). NBE features were assessed through on-site audits and a Geographic Information System. Random forest modeling was applied to estimate NBE features’ predictive capability for CRF, taking into account individual-level characteristics such as sex, BMI z-score, Tanner stage, and moderate to vigorous physical activity (MVPA). This model explained 33% of the variance in the CRF outcome in the out-of-bag test sample. The model identified BMI z-score, sex, and MVPA as the top three predictors, followed by several NBE features, including density of streets with normal traffic, number of intersections, vegetation index, land use mix, building density and social disorder. Overall, the model identified NBE features as having moderate capability to predict CRF. These findings warrant further exploration and could be considered within broader strategies aimed at supporting active living among youth.