The Significance of Local Area Ecologic and Individual-Level Risk Factor Interactions in Understanding Social Determinants of Health

• Activity Number: 410
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Kathryn A. Williams*+ and David N. Williams and Ian McDowell
• Companies: Boston Children's Hospital/University of Ottawa and Boston Children's Hospital/Harvard School of Medicine and University of Ottawa
• Keywords: Ecologic risk factor ; spatial analysis ; hierarchical regression ; geographic hotspots ; adverse birth outcomes ; social determinants of health

Abstract

The ability to link state/province-wide individual data with census aggregate data at local geographic areas allows for granular results and the identification of interactions between ecologic and individual risk factors as these form social determinants of health. Using adverse birth outcomes as a health problem that is only partially predicted from individual risk factors we used two-way interaction tests and contrasts within hierarchical models to explore the relationship between ecologic and individual risk factors. Far too often, subgroup analyses of risk factors have ignored the importance of broader effects that apply across the subgroups, and also the possible heterogeneity among the subgroups. Using data from Ontario, Canada, we show how local area census socio-economic data alter the effect that individual factors such as maternal age and smoking have on birth outcomes. A useful visual representation of interactions advancing on the valuable forest plot was developed. The results suggest that complex spatial and regression analyses of ecologic and individual variables will help inform community-based health intervention efforts.