TWINStudy of Environment, Lifestyle Behaviors, and Health

The role of built (BE) and social environments in supporting healthy lifestyles has received increased attention over the last decade as research and prevention paradigms have shifted from a focus on individual-level behaviour... [ view full abstract ]

The role of built (BE) and social environments in supporting healthy lifestyles has received increased attention over the last decade as research and prevention paradigms have shifted from a focus on individual-level behaviour change to macro-level influences embedded within social-ecological models that affect population health. The TWINS Cohort was set up to link biological and survey data with BE and social environment exposures developed from geocoded residential addresses of adult twins in the Washington State Twin Registry (WSTR). Twin studies enable the detection of environmental effects on health while reducing the structural confounding and biases inherent in studying unrelated individuals who have not been randomly assigned to different environments, as would happen in a randomized controlled trial, which in the case of the BE is neither practical nor ethical under free-living conditions. Twins are thus used as quasi-experimental controls for shared genetic and environmental effect confounds that cannot be held constant via random assignment. We specifically use the twin design to identify causal associations among unique environmental exposures, lifestyle behaviours, and a broad array of physical and mental health outcomes. Subsamples from the TWINS Cohort can be recruited based on select characteristics of interest for detailed, in-person studies. The present study reports on an ongoing prospective sub-cohort of TWINS in which identical (MZ) twins reared together, but who now live apart as adults, wear accelerometers and GPS data loggers for 1-2 weeks to measure objective physical activity levels within a real-time and space continuum. The purpose of this prospective study is to investigate cross-sectional and longitudinal associations between the walkability of the home neighbourhood BE and objective walking and moderate-to-vigorous physical activity (MVPA) levels (minutes per week in activity “bouts”) controlling for genetic and shared environmental influences. The commercially available Walk Score^® index is used to derive a continuous walkability score normalized on a scale of 0-100, with 100 being the most “walkable” neighbourhood. We quantify both walking and MVPA levels inside and outside of empirically defined neighborhoods using the Life Log framework our group previously developed. The home neighborhood BE is quantified within circular buffers of different sizes (e.g., 833m, 1,666m) and types (Euclidean vs. network-based) and is conceptualized using two different definitions based on the location of GPS data points beginning, ending, and/or traversing the pre-defined buffers. In one analysis from this sub-cohort (n=106 individual MZs), objectively measured walking bouts were related to neighborhood walkability, controlling for twin pair ID (random factor), sex, age, BMI, and income (p<0.01). The findings held regardless of buffer size or type (all p<0.05). In ongoing analysis of this sub-cohort, 144 MZ pairs have thus far completed baseline and 15 MZ pairs 3-year follow-up assessments. Mean walking levels within the neighborhood are 60-84 min per week, while MVPA levels within the neighborhood are 35-63 min per week, depending on buffer size and neighborhood definition. Ultimately, this study will determine longitudinal, causal associations between changes in the BE on lifestyle behaviors and health.