Kernel Density Estimation as a Measure of Environmental Exposure Related to Insulin Resistance in Breast Cancer Survivors

• Published in: Cancer epidemiology, biomarkers & prevention biomarkers & prevention, 2017-07, Vol.26 (7), p.1078-1084
• Main Authors: Jankowska, Marta M, Natarajan, Loki, Godbole, Suneeta, Meseck, Kristin, Sears, Dorothy D, Patterson, Ruth E, Kerr, Jacqueline
• Format: Article
• Language: English
• Subjects: Aged; Body Mass Index; Breast cancer; Breast Neoplasms - epidemiology; Breast Neoplasms - pathology; Breast Neoplasms - prevention & control; Built environment; California; Cancer Survivors - statistics & numerical data; Ecological risk assessment; Environmental Exposure - statistics & numerical data; Environmental factors; Exposure; Female; Global positioning systems; GPS; Health risks; Healthy Lifestyle; Humans; Insulin; Insulin - metabolism; Insulin Resistance; Middle Aged; Motor Activity; Neighborhoods; Neoplasm Staging; Recreation; Recreation areas; Regression analysis; Residence Characteristics; Residential density; Spatial Analysis; Urban environments; Walking; Walking - statistics & numerical data
• ISSN: 1055-9965; 1538-7755
• DOI: 10.1158/1055-9965.EPI-16-0927

Environmental factors may influence breast cancer; however, most studies have measured environmental exposure in neighborhoods around home residences (static exposure). We hypothesize that tracking environmental exposures over time and space (dynamic exposure) is key to assessing total exposure. This study compares breast cancer survivors' exposure to walkable and recreation-promoting environments using dynamic Global Positioning System (GPS) and static home-based measures of exposure in relation to insulin resistance. GPS data from 249 breast cancer survivors living in San Diego County were collected for one week along with fasting blood draw. Exposure to recreation spaces and walkability was measured for each woman's home address within an 800 m buffer (static), and using a kernel density weight of GPS tracks (dynamic). Participants' exposure estimates were related to insulin resistance (using the homeostatic model assessment of insulin resistance, HOMA-IR) controlled by age and body mass index (BMI) in linear regression models. The dynamic measurement method resulted in greater variability in built environment exposure values than did the static method. Regression results showed no association between HOMA-IR and home-based, static measures of walkability and recreation area exposure. GPS-based dynamic measures of both walkability and recreation area were significantly associated with lower HOMA-IR ( < 0.05). Dynamic exposure measurements may provide important evidence for community- and individual-level interventions that can address cancer risk inequities arising from environments wherein breast cancer survivors live and engage. This is the first study to compare associations of dynamic versus static built environment exposure measures with insulin outcomes in breast cancer survivors. .