How different data sources and definitions of neighbourhood influence the association between food outlet availability and body mass index: a cross-sectional study

Abstract

Inconsistencies in methodologies continue to inhibit understanding of the impact of the environment on body mass index (BMI). To estimate the effect of these differences we assessed the impact of using different definitions of neighbourhood and datasets on associations between food outlet availability within the environment and BMI. Previous research has not extended to show any differences in the strength of associations between food outlet availability and BMI across both different definitions of neighbourhood and datasets. Descriptive statistics showed differences in the number of food outlets, particularly other food retail outlets between different datasets and definitions of neighbourhood. Despite these differences, our key finding was that across both different definitions of neighbourhood and datasets there was very little difference in size of associations between food outlets and BMI. Researchers should consider and transparently report the impact of methodological choices such as the definition of neighbourhood and acknowledge any differences in associations between the food environment and BMI.

Reconsidering the relationship between fast-food outlets, area-level deprivation, diet quality and body mass index: an exploratory structural equation modelling approach

“…However, previous research using the study sample has shown few differences when using different buffer types and distances. 17 Moreover, our sensitivity analyses (online supplement 4) showed similar associations when using different buffer sizes which are suggested to reflect walking behaviours in the UK. Although buffers were based on the best available evidence, how to define a neighbourhood remains a limitation across the evidence base as it is known individuals may operate beyond a radial buffer, a concept known as the uncertain geographical problem which has been discussed by Kwan extensively.…”

“…Although we acknowledge that individuals are known to operate outside a radial buffer, previous research shows little variation in outcomes by different neighbourhood definitions. 17 Furthermore, when previously using 1600 m radial buffers which are hypothesised to better reflect walking behaviours 18 few differences in associations were seen. 17 Food outlets within each 2 km buffer were counted using a point in polygon analysis using ArcGIS V.10.2.2 (ESRI, Redlands, CA) and summed using a spatial join between food outlet layers and each individual's 2000 m radial buffer.…”

“…17 Furthermore, when previously using 1600 m radial buffers which are hypothesised to better reflect walking behaviours 18 few differences in associations were seen. 17 Food outlets within each 2 km buffer were counted using a point in polygon analysis using ArcGIS V.10.2.2 (ESRI, Redlands, CA) and summed using a spatial join between food outlet layers and each individual's 2000 m radial buffer. Sensitivity analyses were undertaken for 1600 m radial buffers in this study.…”

Food safety and tackling obesity

“…Wadolowska et al 8 present a comparison of dietary patterns derived by two methods and their association with obesity in Polish girls aged 13-21 years as part of the cross-sectional GEBaHealth study, and Mahmood and lowe follow the theme by considering population segmentation as an approach to reducing childhood obesity inequality. 9 it is well reported that childhood obesity is a marker for later life and hence the attention on this age group; however, the food industry has also come under scrutiny in the cross-sectional study by Hobbs et al 10 on how different data sources and definitions of neighbourhood influence the association between food outlet availability and body mass index makes for interesting reading. Finally, the ultimate solution as evaluated by Beaulac and Sandre, 11 also from Canada, could be bariatric surgery which has received the greatest research attention and been associated not only with greater weight loss and health benefit but also with greater risks, complications and, unsurprisingly, financial cost.…”

Investigating the impact of GIS modelled daily exposures to the retail food environment on routinely linked child health data

“…There has been a particular focus on exposure to the 'retail food environment' (RFE) and how this may be linked with obesity. GIS measures of exposure to the RFE widely differ, with studies using different data sources, defining the RFE in different ways, and using heterogeneous 3 methods to measure access or exposure to the RFE [10]. Currently, GIS modelled exposures to the RFE along routes to and from school are not adequate to make reliable predictions about exposure.…”