The goal of this project is to develop an alternative method to conducting field-based audits of the built environment and conduct preliminary validation of the approach's ability to predict physical activity behavior. The mounting number of field audit tools developed in recent years reflects increasing interest in built environment characteristics as determinants of physical activity and related health outcomes. High-resolution omnidirectional imagery (an emerging technology exemplified by Google Street View) provides a potential alternative to field-based approaches for detailed assessment of the built environment. This type of imagery is becoming more widely available and provides the ability to view the built environment from a virtual walk- through or drive-through perspective. Therefore, this study will examine the agreement between omnidirectional imagery and field-based audits. All audits will use the Active Neighborhood Checklist that measures 71 detailed built environment characteristics.
The specific aims are: 1. examine the agreement between detailed measures of the built environment derived from publicly accessible imagery and in-person field audits;and 2. examine the relationship between detailed measures of the built environment derived from publicly available imagery and in-person field audits with physical activity behavior. To measure the agreement between results of field-based audits and those obtained through interpretation of the imagery, we will use observed agreement and Cohen's kappa statistic (K) to compare each checklist item across. Kappa will be used to compare dichotomous items (i.e. yes/no response choice), while the 1-way model intraclass correlation coefficient (ICC) will be used for multiple-choice questions. Because kappa yields higher agreement measures when auditors disagree on the distribution of categories in the data and lower measures when they agree, we also will use Krippendorffer's 1 which does not suffer from this disadvantage. We will use Pearson Product Correlation and T-tests to assess the ability of this approach to predict physical activity behavior. Imagery methods may enhance more comprehensive research on built environment determinants of physical activity and related health outcomes by improving the feasibility of conducting longitudinal studies of changes in the built environment for larger and more geographically dispersed cohorts.
Previous research suggests characteristics of the built environment influence physical activity and related health outcome. The goal of this study is to develop a new method of assessing the built environment that uses innovative image-based technology (e.g. Google Street View). If effective, this new method could advance the field by allowing for multi-regional and/or longitudinal studies of the built environment on physical activity and related health outcomes.
|Kelly, Cheryl; Wilson, Jeffrey S; Schootman, Mario et al. (2014) The built environment predicts observed physical activity. Front Public Health 2:52|
|Kelly, Cheryl M; Wilson, Jeffrey S; Baker, Elizabeth A et al. (2013) Using Google Street View to audit the built environment: inter-rater reliability results. Ann Behav Med 45 Suppl 1:S108-12|