Background:
Although the positive effects of physical activity have been well established, the majority of people remain sedentary. Ecological models state that it is important to understand multiple influences on physical activity to develop appropriate interventions. Demographic and psychosocial correlates have been studied extensively, but physical environmental variables much less. Among physical environmental determinants, neighbourhood ‘walkability’ is often evaluated in the literature. Key elements of a ‘high walkable’ neighbourhood are high street connectivity, high land use mix and high residential density. Recently, several studies in the United States and Australia found a positive relationship between neighbourhood walkability and physical activity in adults. However, in Europe, research on this topic still remains sparse and additional studies on the relationship between walkability and physical activity are certainly necessary.

Objectives:
The purposes of this Belgian study were (1) to investigate whether adults who live in ‘high walkable’ neighbourhoods are more physically active than those living in ‘low walkable’ neighbourhoods, (2) to study whether adults living in high SES neighbourhoods are more physically active than adults who live in low SES neighbourhoods and (3) to investigate whether neighbourhood SES level is a moderator of the relationship between neighbourhood walkability and physical activity.

Methods:
In Ghent, a metropolitan area in Belgium, 16 neighbourhoods were stratified on walkability, based on objectively assessed residential density, connectivity and land use mix. Geographic Information Systems (GIS) were used to measure these parameters. 8 ‘high walkable’ neighbourhoods and 8 ‘low walkable’ neighbourhoods were selected. Also, the neighbourhoods were matched on SES variables, with 8 high SES and 8 low SES neighbourhoods.

Neighbourhood SES level was defined using neighbourhood income data. In total, 800 adults (ages 20-65) participated in the study, 50 per neighbourhood. Data were collected between May 2007 and January 2008. Physical activity levels were assessed using the IPAQ long last seven days interview version. Participants also wore a CSA accelerometer for seven consecutive days and completed the NEWS questionnaire.

Results:
The accelerometer results showed that adults living in a high walkable neighbourhood had significantly more physical activity of moderate intensity (x = 36.2 ± 21.2 min/day) than adults living in a low walkable neighbourhood (x = 29.4 ± 20.6 min/day) (F = 10.66, p < .001). They also reported 48 minutes/week more cycling for transport (F = 16.51, p < .001), 85 minutes/week more walking for transport (F = 59.63, p < .001), 70 minutes/week less use of motorized transport (F = 10.78, p < .001) and 15 minutes/week more walking for recreation (F = 4.75, p = .03).

When analysing SES differences in physical activity, accelerometer results showed that adults living in a low SES neighbourhood had significantly more physical activity of moderate intensity (x = 34.5 ± 22.4 min/day) than those living in a high SES neighbourhood (x = 21.2 ± 19.8 min/day) (F = 5.41, p = .02). They also reported 45 minutes/week more walking for transport (F = 22.16, p < .001) and 80 minutes/week less use of motorized transport (F = 15.95, p < .001). The amount of cycling for transport did not differ between inhabitants of the high SES and low SES neighbourhoods.

In answer to the final research question, neighbourhood SES level did not significantly moderate any relationship between walkability and physical activity behaviour.

Conclusions:
In this Belgian study, previous findings of studies in Australia and the United States were partially confirmed. Prior studies showed that physical environmental factors are mainly associated with physical activity for transport. A new finding in this study was the association between neighbourhood walkability and walking for recreation.

The results on the relation between neighbourhood SES and physical activity were not in line with previous studies, which found that living in a high SES neighbourhood was associated with more physical activity. In this study, adults living in low SES neighbourhoods were more physically active. Possible explanations for this finding are that inhabitants of low SES neighbourhoods are more physically active at work and are less likely to possess a car.
In this study, a strong association between walkability and physical activity was established. Therefore, the findings of this study are very promising within the scope of future environmental interventions and policy initiatives to increase physical activity. However, additional longitudinal studies are certainly needed to determine causal relations.

Support:
Thanks to the Ghent University supporting the grant for Delfien Van Dyck.