Patterning of neighbourhood food outlets and longitudinal associations with children's eating behaviours.

This study examined cross-sectional and prospective associations between typologies of neighbourhood food environment and dietary patterns among 10-12 year-old children. Baseline data were collected in 2003 and follow-up data in 2006 from children in Melbourne or Geelong. Parents completed a food frequency questionnaire at both time points. 'Healthful' and 'energy-dense' dietary pattern scores were computed. A Geographic Information System was used to determine the presence or absence of food outlets (cafés/restaurant; fast food; supermarkets/grocery stores; convenience store; greengrocer; and butcher, seafood or poultry retailer) within an 800 m road network buffer of home. Three typologies were identified: 1-variety of food outlets, including those selling core/fresh foods (n = 96); 2-café/restaurant and convenience (n = 160); 3-few types of outlets (n = 208). Latent class analysis was used to identify underlying unobservable typologies of neighbourhood food outlet availability. Linear mixed models were fitted to determine cross-sectional (n = 439) and longitudinal (n = 173) associations between the three identified neighbourhood typologies and each (log-transformed) dietary pattern, accounting for clustering within families and schools. There was little evidence of cross-sectional associations. The longitudinal analyses showed that compared to those with a variety of food outlets, those with few types had 25% lower scores for the healthful dietary pattern (p < 0.05) three years later. For optimal dietary patterns, availability of a variety of food outlets close to home, particularly those where core/fresh foods are available, may be important.