Defining neighborhood walkability for children in the ELFE cohort study

Abstract Background Although walkability is known to be associated with obesity in adults, there is a paucity of data evaluating the association of walkability and health outcomes in children. In particular, little is known regarding the possible association with children’s mental health. Methods We used data from the Etude Longitudinale Française depuis l’Enfance (ELFE) cohort study. Participating mothers gave birth in 2011 in a representative sample of 320 maternity hospitals in mainland France. This pilot study is limited to participants in Paris and children at 5.5 years old. Walkability is defined by three parameters: population density (defined by census data), street connectivity (defined by intersection density), and land-use mix (entropy index). To approximate walkable areas, 500-meter street-network buffer zones were created around each participants address using a geographic information system. Children’s mental health was assessed using parent-reported Strengths and Difficulties Questionnaire at age 5.5 years. Results There was a total of 280 children in this pilot study. On measures of land-use mix, we found a mean of 0.79 (SD: 0.12), representing high accessibility. Residential and intersection density were highly correlated (p < 0.01). The walkability index demonstrated high variability among participants (mean: 10.6; SD: 10.44; range: -1.67 - 30.59). Conclusions Developing a walkability index is a multifactorial measure to describe a children’s neighborhood environment. Future work includes determining the association with children’s mental health disorders, while accounting for individual and neighborhood variables. This ecological model can inform social epidemiology and contextual disparities in children’s health. Key messages • Little is known about neighborhood walkability and the association with children’s mental health. • Using objective measures to calculate neighborhood composition and help elucidate this relationship.