A cross‐sectional and prospective examination of alcohol use and misuse among adult twin and sibling pairs discordant for neighborhood socio‐economic disadvantage
Wendy S. Slutske, Timothy B. Baker, Thomas M. PiaseckiAbstract
Aim
To determine whether the association between neighborhood disadvantage and alcohol use and misuse can be explained by a potentially causal effect of neighborhoods or by between‐family genetic and environmental confounding.
Design
Cross‐sectional and prospective discordant twin and sibling study.
Setting
United States national sample.
Participants
1477 adult twin and sibling pairs (217 monozygotic twin, 336 dizygotic twin and 924 full biological sibling pairs) from the National Longitudinal Study of Adolescent to Adult Health (mean age = 28.49 years, 53% female). Most twin and sibling pairs were discordant for the level of disadvantage in their neighborhood (70%).
Measurements
The predictor was wave 4 neighborhood disadvantage, along with wave 4 individual‐, family‐ and neighborhood‐level covariates. The outcomes were any alcohol use, quantity x frequency of alcohol use (alcohol QF) and frequency of binge drinking assessed at wave 4 in the cross‐sectional models and at wave 5 in the prospective models. Neighborhood disadvantage for the two members of a twin/sibling pair was used to create two orthogonal within‐ and between‐pair variables to be used to decompose the neighborhood/alcohol association into a within‐pair (i.e. potentially causal) effect and a between‐pair association (i.e. family level confounding).
Findings
In fully adjusted models including all covariates, only between‐pair (not within‐pair) neighborhood disadvantage was cross‐sectionally [odds ratio (OR) = 0.92, 95% confidence interval (CI) = 0.88–0.98] and prospectively (OR = 0.92, 95% CI = 0.85–0.98) associated with lower odds of any alcohol use in the past year. Between‐pair neighborhood disadvantage was cross‐sectionally, but not prospectively, associated with increased alcohol QF in the past month among women [mean ratio (MR) = 1.09, 95% CI = 1.03–1.15], but not among men (MR = 1.01, 95% CI = 0.96–1.06). Within‐pair disadvantage was also cross‐sectionally, but not prospectively, associated with increased alcohol QF in the past month (MR = 1.04, 95% CI = 1.01–1.07). Neighborhood disadvantage was not significantly associated with binge drinking.
Conclusions
Neighborhood disadvantage appears to be negatively, positively and not significantly associated with any alcohol use, quantity‐frequency of use and binge drinking, respectively. Neighborhood disadvantage associations are more consistently explained by between‐family confounds, that is, genetic and family environmental factors common to neighborhood disadvantage and to drinking occurrence and amount consumed. This may be due to differential selection by families into neighborhoods.