Objective: To examine whether compositional and/or contextual area characteristics are associated with area socioeconomic inequalities and between-area differences in recreational cycling. Setting: The city of Melbourne, Australia. Participants: 2349 men and women residing in 50 areas (58.7% response rate). Main outcome measure: Cycling for recreational purposes (at least once a month vs never). Design: In a cross-sectional survey participants reported their frequency of recreational cycling. Objective area characteristics were collected for their residential area by environmental audits or calculated with Geographic Information Systems software. Multilevel logistic regression models were performed to examine associations between recreational cycling, area socioeconomic level, compositional characteristics (age, sex, education, occupation) and area characteristics (design, safety, destinations or aesthetics). Results: After adjustment for compositional characteristics, residents of deprived areas were less likely to cycle for recreation (OR 0.66; 95% CI 0.43 to 1.00), and significant between-area differences in recreational cycling were found (median odds ratio 1.48 (95% credibility interval 1.24 to 1.78). Aesthetic characteristics tended to be worse in deprived areas and were the only group of area characteristics that explained some of the area deprivation differences. Safety characteristics explained the largest proportion of between-area variation in recreational cycling. Conclusion: Creating supportive environments with respect to safety and aesthetic area characteristics may decrease between-area differences and area deprivation inequalities in recreational cycling, respectively.

doi.org/10.1136/jech.2007.067116, hdl.handle.net/1765/13618
Journal of Epidemiology and Community Health
Erasmus MC: University Medical Center Rotterdam

Kamphuis, C., Giskes, K., Kavanagh, A. M., Thornton, L. E., Thomas, L., van Lenthe, F., … Turrell, G. (2008). Area variation in recreational cycling in Melbourne: a compositional or contextual effect?. Journal of Epidemiology and Community Health, 62(10), 890–898. doi:10.1136/jech.2007.067116