Smoking prevalence of Asian Americans: Does living in an ethnic enclave matter?

PR-5

The ethnic composition of a neighborhood can affect smoking-related behaviors through a variety of causal pathways including social support, stress, cultural norms, modeling of smoking behavior, neighborhood economic structure, availability and advertising of cigarettes, and health information. Little is known about how neighborhood-level factors affect smoking in Asian Americans.
 The objective of this study was to determine whether neighborhood ethnic composition, individual perceptions of neighborhood social cohesion, or neighborhood socioeconomic status (SES) were associated with smoking prevalence in a population-based sample of Asian Americans, independent of individual factors. We hypothesized that living in a neighborhood with a higher percentage of Asians would be associated with higher rates of smoking in Asian American men and lower rates of smoking in Asian American women, mirroring the smoking norms and practices in most Asian countries.
 Individual-level data, including smoking, age, gender, race/ethnicity, marital status, education, poverty status, employment, percent of life in the US, language spoken at home, and perceived neighborhood social cohesion (a scale tapping the extent of connectedness, trust, and solidarity among neighbors; coefficient of alpha=0.73), were obtained or constructed from the 2003 California Health Interview Survey (CHIS). CHIS is a cross-sectional, population-based telephone survey of 42,000 civilian households in California. A neighborhood was defined as the census tract and participants’ census tracts were linked to data from the 2000 Census. The ethnic composition of the neighborhood was defined as the proportion in the census tract who were Asian. Neighborhood SES was constructed using principal component factor analysis with orthogonal rotation from four Census measures that are highly correlated: concentrated affluence, concentrated poverty, % of college-educated residents, and % of house ownership. The reliability coefficient was 0.83. Smoking was dichotomized as current smoking or not. Gender-stratified multiple regression models with robust variance estimates were used to account for correlations among residents of the same neighborhood.
 The sample included 1693 Asian men and 2174 Asian women, ages 18 and older: 22% of Asian men and 6% of Asian women were current smokers. For Asian women, an increasing proportion Asians in the tract was significantly associated with lower adjusted odds of smoking (OR=0.14, 95% CI=0.03,0.80), independent of age, marital status, individual SES, percent of life in the US, language spoken at home, perceived neighborhood social cohesion, and neighborhood SES. For men, the proportion Asians in the neighborhood was not associated with smoking; however, higher levels of perceived neighborhood social cohesion were independently associated with lower odds of smoking (OR=0.74, 95% CI=0.61, 0.90). Neighborhood SES was not significant for men or women.
 For Asian American women, ethnic composition, which may represent social norms and modeling of behavior, may mediate smoking. For men, a sense of being disconnected and distrustful of their surroundings (a possible marker of stress or low social support), may mediate smoking behavior. Neighborhood structural position, typically measured by SES, is not necessarily relevant for all ethnic groups or health outcomes