Influence of neighbourhood ethnic density, diet and physical activity on ethnic differences in weight status: A study of 214,807 adults in Australia

We investigated whether ethnic and country of birth differences in adult Body Mass Index (BMI) were associated with differences in diet, physical activity and ethnic density (the percentage of an ethnic group within the neighbourhood environment). A sample of 214,807 adults living in Australia was extracted from the 45 and Up Study. Analyses comprised multilevel modelling of BMI for 38 ethnic and country of birth groups. Physical activity was ascertained using the Active Australia Survey. Dietary measures included self-reported consumption of fruit, vegetables, meat and cheese. Ethnic density was objectively measured using 2006 Australian Census data. Possible confounders included age, gender, household income, educational qualifications, economic status, couple status, language, duration of residence, neighbourhood affluence and remoteness. Compared to Australian-born Australians (age-gender adjusted mean BMI = 27.1, 95%CI 27.1, 27.2), overseas-born groups often had lower mean BMI, especially the Chinese born in China (23.2, 23.0, 23.4). Exceptions included the Italians (BMI = 28.1), Greeks (28.5), Maltese (27.6), Lebanese (28.4) and Croatians (27.8) born in their ethnic-country of origin. Regardless of birthplace, BMI was lower for the English, Scottish, and Chinese, but higher for Italians and Greeks. Some ethnic differences reflected the 'healthy migrant' hypothesis, whereas others did not. These differences were only partially attenuated by controls for portions of fruit and vegetables, meat and cheese, frequency of participation in physical activity, and other explanatory variables. Ethnic density was associated with lower BMI for the English and Irish (p < 0.05), regardless of whether they were born in the UK, Ireland, or Australia. Ethnic differences in adult weight status in Australia do not appear to be fully explained by conventional risk factors. For some groups, but not all, living among others of the same ethnic group may proxy unmeasured health-promoting factors and these contexts, along with other factors that harm health (e.g. racial discrimination) warrant further investigation. (C) 2013 Elsevier Ltd. All rights reserved.