Objective: To explore socio-economic differences in use of staff canteens and whether frequent use of staff canteens is associated with different food patterns and obesity. Design: Cross-sectional study using three self-administered questionnaires, two of them including food frequency questions. Factor analysis was used to explore food patterns. Setting: Oslo, Norway, 2000-2001. Subjects: In total 8943 adult, working Oslo citizens. Results: Frequent ($3 times/week) use of staff canteens was most likely among men, younger workers and those in the highest education and income groups. However, after adjustment for demographic, socio-economic and lifestyle factors, those with highest education were least likely to use staff canteens frequently. Frequent eating in staff canteens was positively associated with a Western food pattern (based on fat-rich food, fast food and red meat) and inversely associated with a traditional food pattern (based on boiled potatoes and gravy, and less rice, pasta and oil) in multivariate analyses. Unadjusted, frequent eating in staff canteens was also inversely associated with a prudent food pattern (based on fruit, vegetables, fish, legumes and oil). The likelihood of being obese (BMI $ 30 kg/m 2 ) increased significantly with frequent eating in staff canteens, also when adjusted for demographic and socio-economic variables. Adjustment for the food patterns attenuated this relationship, but it was still significant. Conclusions: Frequent eating in staff canteens was negatively related to socioeconomic position and positively associated with unhealthy dietary habits. This partly explained higher odds for obesity among frequent users of staff canteens. Future research should assess the availability and food options of staff canteens.

Socioeconomic differences in overweight are well documented, but most studies have only used one or two indicators of socioeconomic position (SEP). The aim of the present study was to explore the relative importance of indicators of SEP (occupation, education and income) in explaining variation in BMI and waist:hip ratio (WHR), and the mediating effect of work control and lifestyle factors (dietary patterns, smoking and physical activity). The Oslo Health Study, a cross-sectional study, was carried out in 2000-1, Oslo, Norway. Our sample included 9235 adult working Oslo citizens, who attended a health examination and filled in two complementary FFQ with , 20 % missing responses to food items. Four dietary patterns were identified through factor analysis, and were named 'modern', 'Western', 'traditional' and 'sweet'. In multivariate models, BMI and WHR were inversely associated with education (P,0·001/P, 0·001) and occupation (P¼ 0·002/P,0·001), whereas there were no significant associations with income or the work control. The 'modern' (P, 0·001) and the 'sweet' (P,0·001) dietary patterns and physical activity level (P, 0·001) were inversely associated, while the 'Western' dietary pattern (P,0·001) was positively associated with both BMI and WHR. These lifestyle factors could not fully explain the socioeconomic differences in BMI or WHR. However, together with socioeconomic factors, they explained more of the variation in WHR among men (21 %) than among women (7 %).