Vivienne C. Ivory, Tony Blakely, Ken Richardson, George Thomson, Kristie Carter

Health behavior takes place within social contexts. In this study, we investigated whether changes in exposure to neighborhood deprivation and smoking prevalence and to household smoking were associated with change in personal smoking behavior. Three waves of biannual data collection (2004-2009) in a New Zealand longitudinal study, the Survey of Family, Income and Employment (SoFIE)-Health, were used, with 13,815 adults (persons aged ≥15 years) contributing to the analyses. Smoking status was dichotomized as current smoking versus never/ex-smoking. Fixed-effects regression analyses removed time-invariant confounding and adjusted for time-varying covariates (neighborhood smoking prevalence and deprivation, household smoking, labor force status, income, household tenure, and family status). A between-wave decile increase in neighborhood deprivation was significantly associated with increased odds of smoking (odds ratio (OR) = 1.08, 95% confidence interval (CI): 1.02, 1.14), but a between-wave increase in neighborhood smoking prevalence was not (OR = 1.04, 95% CI: 0.98, 1.10). Changing household exposures between waves to live with another smoker (compared with a nonsmoker (referent)) increased the odds of smoking (OR = 2.48, 95% CI: 1.84, 3.34), as did changing to living in a sole-adult household (OR = 1.52, 95% CI: 1.07, 2.14). Tobacco control policies and programs should address the broader household and neighborhood circumstances within which individual smoking takes place.