Creating a replicable, valid cross-platform buffering technique: The sausage network buffer for measuring food and physical activity built environments

BACKGROUND Obesity researchers increasingly use geographic information systems to measure exposure and access in neighborhood food and physical activity environments. This paper proposes a network buffering approach, the "sausage" buffer. This method can be consistently and easily replicated across software versions and platforms, avoiding problems with proprietary systems that use different approaches in creating such buffers. METHODS In this paper, we describe how the sausage buffering approach was developed to be repeatable across platforms and places. We also examine how the sausage buffer compares with existing alternatives in terms of buffer size and shape, measurements of the food and physical activity environments, and associations between environmental features and health-related behaviors. We test the proposed buffering approach using data from EAT 2010 (Eating and Activity in Teens), a study examining multi-level factors associated with eating, physical activity, and weight status in adolescents (n=2,724) in the Minneapolis/St. Paul metropolitan area of Minnesota. RESULTS Results show that the sausage buffer is comparable in area to the classic ArcView 3.3 network buffer particularly for larger buffer sizes. It obtains similar results to other buffering techniques when measuring variables associated with the food and physical activity environments and when measuring the correlations between such variables and outcomes such as physical activity and food purchases. CONCLUSIONS Findings from various tests in the current study show that researchers can obtain results using sausage buffers that are similar to results they would obtain by using other buffering techniques. However, unlike proprietary buffering techniques, the sausage buffer approach can be replicated across software programs and versions, allowing more independence of research from specific software.