Abstract
Background
Underage binge drinking is a serious health concern that is likely influenced by the neighbourhood environment. However, longitudinal evidence has been limited and few studies have examined time-varying neighbourhood factors and demographic subgroup variation.
Methods
We investigated neighbourhood influences and binge drinking in a national cohort of US 10th grade students at four times (2010–2014; n = 2745). We estimated odds ratios (OR) for past 30-day binge drinking associated with neighbourhood disadvantage, personal and property crime (quartiles), and number of liquor, beer and wine stores within 5 km, and then evaluated whether neighbourhood associations differ by age, sex and race/ethnicity.
Results
Neighbourhood disadvantage was associated with binge drinking before 18 [OR = 1.54; 95% confidence interval (1.14, 2.08)], but not after 18 years of age. Property crime in neighbourhoods was associated with a higher odds of binge drinking [OR = 1.54 (0.96, 2.45)], an association that was stronger in early adulthood [4th vs 1st quartile: OR = 1.77 (1.04, 3.03)] and among Whites [4th vs 1st quartile: OR = 2.46 (1.03, 5.90)]. Higher density of liquor stores predicted binge drinking among Blacks [1–10 stores vs none: OR = 4.31 (1.50, 12.36)] whereas higher density of beer/wine stores predicted binge drinking among Whites [one vs none for beer: OR = 2.21 (1.06, 4.60); for wine: OR = 2.04 (1.04, 4.03)].
Conclusions
Neighbourhood conditions, particularly those related to economic circumstances, crime and alcohol outlet density, were related to binge drinking among young adults, but associations varied across age and individual characteristics.
Keywords: Alcohol, binge, crime, disadvantage, neighbourhood, outlets
Key Messages
Living in disadvantaged neighbourhoods is associated with binge drinking in late adolescence, but this relationship did not persist into adulthood.
Personal crime was associated with less binge drinking, but higher property crime increased the chance of binge drinking in early adulthood and among Whites.
Binge drinking in Whites was related to living near more beer and wine stores, whereas Blacks were more likely to binge drink living near more liquor stores.
Introduction
Neighbourhood environments are believed to play an important role in adolescent development and trajectories of health into adulthood.1,2 Health-related behaviours such as alcohol use may be sensitive to environmental influences.3 Alcohol use develops from adolescence through early adulthood and can lead to heavy or problematic drinking that carries significant acute and long-term risks for physical and psychological harm.4,5 Therefore, determining how neighbourhood environments influence drinking development during adolescence could provide important insights into preventing alcohol-related harms.
Socio-economic disadvantage and social disorganization are two features of neighbourhoods believed to influence alcohol use. Disadvantage is characterized by rates of poverty, unemployment, school drop-out and single parent families. Social disorganization is characterized by weak social ties within the community, poor collective efficacy, and deficiency in informal social controls that contribute to crime and deviant youth behaviours like drinking.6 Drinking could be a means of coping with stress and depression from living in environments with poor educational and employment opportunities, violence and crime.7–9 These neighbourhoods may also be deficient in institutional resources (e.g. police, parks, libraries, community services) that promote safe and healthy development.10
Alcohol availability in the community also influences alcohol consumption. A higher density of outlets makes alcohol more physically accessible, reduces costs through competition and signals reduced social stigma towards drinking.11,12 Outlets are frequently concentrated in disadvantaged, high-crime neighbourhoods and are believed to attract crime and violence.13,14 Studies on alcohol outlets also have to consider differences between on-premises (e.g. bars, nightclubs, restaurants) or off-premises (e.g. grocery and liquor stores) consumption and the types of alcohol sold (e.g. hard liquor, beer and wine).15
Recent reviews have reported mixed or contradictory evidence of how neighbourhood disadvantage, social disorganization and availability affect drinking. Three reviews reported that both disadvantage and advantage were associated with drinking.16–18 One review found that social disorganization and crime were associated with either a higher risk of drinking or no difference after accounting for other factors.17 One review concluded that alcohol outlet density had little impact on individual drinking,15 whereas another found suggestive, but inconclusive evidence that it increased drinking, especially among adolescents.19 All reviews stressed the need for additional longitudinal studies as reviews were based largely on cross-sectional or ecological data. Reviews also speculated whether mixed findings may be due in part to neighbourhoods having different effects depending upon an individual’s age, sex or race/ethnicity. More recent studies using longitudinal cohorts generally report positive associations between neighbourhood disadvantage and drinking outcomes and variations by sex and race.20–23 However, none of these studies examined all three neighbourhood factors together over multiple time points throughout the important period from late adolescence into early adulthood.
Accordingly, we used data from a recent, nationally representative cohort study of US 10th-grade students followed into early adulthood to address the following questions. (i) To what extent are neighbourhood disadvantage, crime and the number of alcohol outlets associated with binge drinking? (ii) Is the relationship between neighbourhood context and binge drinking modified by age, sex or race/ethnicity? (iii) To what degree do these relationships vary by the type of off-premises alcohol outlet (i.e. liquor, beer or wine)?
Methods
Sample
The NEXT Generation Health Study is a nationally representative cohort study of U.S. 10th-grade students recruited during the 2009–2010 academic year (Wave I). A multi-stage stratified probability sampling design was used to select school districts and schools within districts with the probability of selection proportional to enrollment. All students within 1−5 randomly selected 10th-grade classes in each school were invited to participate. A total of 81 schools participated and 2784 10th-graders completed in-school surveys. Six annual follow-up surveys were conducted (Waves II−VII) either in school or online with ≥78.2% participation rates. We used data for those waves in which participants’ geocoded home addresses were available to be linked to neighbourhood data: Wave I [Time 1 (T1), 10th grade], Wave III [Time 2 (T2), 12th grade], Wave IV [Time 3 (T3), 1 year after high school], and Wave V [Time 4 (T4), 2 years after high school] (n = 2745). There was intermittent loss to follow-up: 61% of the total sampled participated in all four time points, 81% participated in three or more and 92% participated in 2 or more. All participants provided informed consent and the study protocol was approved by the institutional review board (IRB) for the Eunice Kennedy Shriver National Institute of Child Health and Human Development.
Measures
Binge drinking
Past 30-day binge drinking was defined as consuming five (for boys) or four (for girls) or more drinks either in a row or on an occasion within 2 h (yes vs no). A ‘drink’ was defined as a can or bottle of beer, a glass of wine or a wine cooler, a shot of liquor, or a mixed drink with liquor in it.
Neighbourhood disadvantage and demographics
Five-year average estimates from 2006–2010, 2008–2012, 2009–2013 and 2010–2014 American Community Survey (ACS) were used to create a neighbourhood disadvantage score for each US census tract based on the following 10 indicators: median income; families below 100% of federal poverty level; use of food stamps; female-headed households with children under 18 years; Gini coefficient (a measure of income inequality); residents with high school education or greater; bachelor’s degree or greater; male unemployment; female unemployment; and white-collar workers. Indicators were standardized and then summed into a standardized score of disadvantage. Scores were then linked to participants’ data via their geocoded census tracts for the closest time period. Neighbourhood median age, the proportion of Black residents and population density measured from the ACS were considered as potential confounders.
Neighbourhood crime and alcohol outlets
Indices of property (burglary, motor vehicle theft) and personal crime (murder, rape and assault) at the census block level were obtained from Applied Geographic Solutions (AGS). Each index was standardized to a national average of 100. Indices weighted each category of crime equally (e.g. murder was weighted no differently than rape or assault). Due to positive skew, we categorized crime indices into quartiles. Separate counts of liquor, beer and wine stores within 5 km of participants’ homes were available at T1 and T3 only. Therefore, T1 values were carried forward for T2, and T3 values were carried forward for T4. Alcohol outlet measures were based on business address data provided by Dun & Bradstreet (www.dnb.com). Values were categorized into quartiles for liquor stores (none, 1–10 stores, 11–20 stores, and ≥21 stores) and tertiles for beer and wine stores (0, 1, ≥2).
Participant demographics
Age, sex, race/ethnicity (White, Black, Hispanic or Other), family affluence and family composition at T1 were considered as potential confounders. Family composition was defined as (i) both biological parents, (ii) one biological parent and a stepparent/partner, (iii) single parent, or (iv) other (e.g. grandparent). Family affluence was based on self-reported number of computers in the household, whether respondents had their own bedrooms, family automobile ownership and family vacations taken within the last year.24
Analyses
Multilevel mixed-effects logistic regression was used to estimate the odds of binge drinking associated with disadvantage, crime and liquor stores. Random intercepts were included to account for within-school and within-individual clustering. We estimated unadjusted and then adjusted models controlling for neighbourhood and participant demographic factors. To estimate subgroup heterogeneity (effect modification) we added neighbourhood by age, sex and race/ethnicity interactions to the model and then used linear combinations of model coefficients to produce stratified estimates. Estimates accounted for sampling weights and were reported as odds ratios (OR) with corresponding 95% confidence intervals (CI). Analyses were conducted in Stata version 14.2.25
Results
Individual and neighbourhood characteristics
The mean age of the cohort was 16.2 years at T1 and 20.2 years at T4, 45.6% were male, 57.9% White, 14.9% Black and 19.7% Hispanic (Table 1). Half (49.1%) were from families of moderate affluence, half (52.2%) lived with both biological parents, and 24.3% came from single-parent families. Binge drinking increased from 27.7% to 41.5% over time. Mean neighbourhood disadvantage at T1 was 0.1 (± 0.2) indicating the cohort represented neighbourhoods one-tenth of a standard deviation higher than the national average; disadvantage did not appreciably change over time. Mean neighbourhood personal and property crime increased from 83.3 to 94.0 and 78.8 to 92.6, respectively. This indicated that the cohort tended to live in areas of less crime than the national average (100.0), but exposure increased over time. The mean number of liquor (5.5), beer (1.7) and wine (0.7) stores within 5 km changed little over time from T1.
Table 1.
Individual and neighbourhood characteristics in the NEXT Generation Health Study, 2010–2014a
| Characteristic | Time 1 (n = 2465) |
Time 2 (n = 2363) |
Time 3 (n = 2113) |
Time 4 (n = 2121) |
||||
|---|---|---|---|---|---|---|---|---|
| Mean | ± | Mean | ± | Mean | ± | Mean | ± | |
| Individual level | ||||||||
| Past 30-day binge drinking (%) | 27.7 | 3.6 | 27.3 | 3.9 | 40.7 | 5.3 | 41.5 | 5.8 |
| Age (years) | 16.2 | 0.04 | 18.2 | 0.1 | 19.1 | 0.04 | 20.3 | 0.04 |
| Male (%) | 45.6 | 3.2 | 44.7 | 3.2 | 41.3 | 4.0 | 39.2 | 3.9 |
| Race/ethnicity (%) | ||||||||
| White | 57.9 | 10.5 | 58.7 | 11.7 | 62.1 | 11.4 | 61.3 | 10.6 |
| Black | 14.9 | 6.6 | 14.8 | 7.6 | 11.6 | 6.1 | 11.7 | 6.2 |
| Other/multiracial | 7.6 | 2.4 | 6.6 | 2.0 | 6.9 | 2.1 | 7.6 | 2.3 |
| Hispanic | 19.7 | 7.5 | 19.9 | 7.6 | 19.4 | 8.5 | 19.4 | 7.7 |
| Family affluence (%) | ||||||||
| Low | 23.8 | 5.3 | 23.0 | 6.0 | 22.3 | 5.3 | 22.4 | 5.2 |
| Moderate | 49.1 | 2.9 | 49.0 | 2.8 | 48.3 | 3.5 | 49.1 | 2.8 |
| High | 27.1 | 4.9 | 28.0 | 5.5 | 29.4 | 5.3 | 28.5 | 5.3 |
| Family composition (%) | ||||||||
| Both parents | 52.2 | 4.4 | 51.2 | 5.0 | 53.7 | 5.4 | 52.3 | 5.2 |
| One parent and stepparent | 19.5 | 2.5 | 16.3 | 2.4 | 16.5 | 3.1 | 15.9 | 2.3 |
| Single parent | 24.3 | 3.6 | 25.5 | 5.3 | 23.7 | 5.0 | 24.5 | 5.3 |
| Other | 4.0 | 1.1 | 7.0 | 1.9 | 6.2 | 1.9 | 7.4 | 2.3 |
| Neighbourhood level | ||||||||
| Socio-economic disadvantage | 0.1 | 0.2 | 0.2 | 0.2 | 0.2 | 0.1 | 0.1 | 0.2 |
| Personal crime | 83.3 | 30.1 | 93.8 | 37.4 | 97.3 | 27.5 | 94.0 | 30.7 |
| Property crime | 78.8 | 20.3 | 83.6 | 20.0 | 91.0 | 15.0 | 92.6 | 19.8 |
| Number of beer stores within 5 km | 1.7 | 1.6 | 1.6 | 1.5 | 1.7 | 1.1 | 1.9 | 1.3 |
| Number of wine stores within 5 km | 0.7 | 0.9 | 0.7 | 0.8 | 0.9 | 0.7 | 0.9 | 0.7 |
| Number of liquor stores within 5 km | 5.5 | 5.1 | 5.9 | 5.1 | 6.6 | 4.3 | 6.5 | 4.2 |
| Median age (years) | 37.7 | 1.4 | 38.2 | 1.6 | 34.0 | 1.2 | 36.9 | 1.2 |
| Male (%) | 48.9 | 0.7 | 49.0 | 0.6 | 49.1 | 0.4 | 49.2 | 0.4 |
| Black (%) | 10.5 | 4.7 | 12.6 | 7.0 | 12.1 | 4.6 | 11.5 | 4.9 |
| Population density (per square mile) | 3927.5 | 3085.4 | 4358.8 | 2991.4 | 5329.5 | 2432.8 | 5073.4 | 2318.5 |
Estimates weighted to the population; ± = 1.96 x standard error.
Neighbourhood contexts associated with binge drinking
Neither neighbourhood disadvantage nor the number of liquor stores within 5 km was associated with binge drinking in unadjusted or adjusted models (Table 2). Living in neighbourhoods with the highest quartile of property crime was associated with binge drinking [4th vs 1st quartile: OR = 1.70; 95% CI (1.01, 2.87)], but this association was attenuated in the adjusted model [OR = 1.54; 95% CI (0.96, 2.45)]. Living in the 2nd compared with the lowest quartile of personal crime was associated with less binge drinking [OR = 0.58; 95% CI (0.39, 0.88)] and was similar to the 4th quartile estimate. School- (0.53) and individual-level (3.9) random intercepts translated into 7% and 58%, respectively, of the total variance being accounted for by clustering.
Table 2.
Odds of past 30-day binge drinking associated with neighbourhood socio-economic disadvantage, crime and liquor stores. NEXT Generation Health Study, 2010–2014 (n = 2, 745)
| Characteristic | Unadjusted |
Adjusteda |
||
|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | |
| Fixed effects | ||||
| Socioe-conomic disadvantage | 1.23 | (0.98, 1.54) | 1.05 | (0.88, 1.25) |
| Personal crime | ||||
| 1st quartile | 1.00 | Referent | 1.00 | Referent |
| 2nd quartile | 0.89 | (0.55, 1.42) | 0.58 | (0.39, 0.88) |
| 3rd quartile | 1.17 | (0.69, 2.00) | 0.70 | (0.43, 1.13) |
| 4th quartile | 1.12 | (0.66, 1.88) | 0.59 | (0.36, 0.97) |
| Property crime | ||||
| 1st quartile | 1.00 | Referent | 1.00 | Referent |
| 2nd quartile | 1.00 | (0.62, 1.62) | 0.87 | (0.57, 1.34) |
| 3rd quartile | 1.05 | (0.65, 1.69) | 1.07 | (0.67, 1.71) |
| 4th quartile | 1.70 | (1.01, 2.87) | 1.54 | (0.96, 2.45) |
| Number of liquor stores in 5 km | ||||
| None | 1.00 | Referent | 1.00 | Referent |
| 1 − 10 stores | 1.32 | (0.82, 2.14) | 1.09 | (0.67, 1.78) |
| 11 − 20 stores | 1.25 | (0.58, 2.69) | 1.04 | (0.55, 1.98) |
| ≥21 | 1.28 | (0.56, 2.91) | 1.10 | (0.57, 2.13) |
| Random effects | ||||
| School-level intercept | NAb | 0.53 | (0.23, 1.24) | |
| Individual-level intercept | NAb | 3.93 | (2.46, 6.27) | |
Confidence interval (CI); adjusted odds ratio (aOR); NA, not applicable.
Controlling for neighbourhood (median age, % Black, population density) and individual factors (age, sex, race/ethnicity, family affluence, family composition).
Random effects varied across unadjusted models; school-level intercept ranged from 0.77 to 0.87 and individual-level intercept ranged from 3.10 to 3.15.
Neighbourhood associations by age, sex and race/ethnicity
Interaction tests of age with disadvantage [χ2 (1) = 12.50; P < 0.01] and personal crime [χ2 (3) = 9.59; P = 0.02] suggested that neighbourhood associations varied over time irrespective of sex and race/ethnicity (Table 3). A standard deviation difference in disadvantage was associated with a 54% higher odds of binge drinking before age 18 [OR = 1.54; 95% CI (1.14, 2.08)] but not after age 18 [OR = 0.92; 95% CI (0.77, 1.09)]. By contrast, personal crime was inversely associated with binge drinking after age 18 only [2nd vs 1st quartile: OR = 0.52; 95% CI (0.33, 0.80); 4th vs 1st quartile: OR = 0.48; 95% CI (0.27, 0.87)]. Neighbourhood property crime was positively associated with binge drinking [4th vs 1st quartile: OR = 1.77; 95% CI (1.04, 3.03)], but evidence of heterogeneity was weak [χ2 (3) = 2.33; P = 0.5061]. Density of liquor stores was not associated with binge drinking across age [χ2 (3) = 5.77; P = 0.1231].
Table 3.
Adjusted odds of past 30-day binge drinking associated with neighbourhood socio-economic disadvantage, crime and liquor stores by age, sex and race. NEXT Generation Health Study, 2010–2014 (n = 2, 745)
| Characteristic | By ageab |
By sexa,b |
By race/ethnicityab |
|||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Under 18 |
18 or Older |
Female |
Male |
White |
Black |
Other |
Hispanic |
|||||||||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | |
| Fixed effects | ||||||||||||||||
| Socio-economic disadvantage | 1.54 | (1.14, 2.08) | 0.92 | (0.77, 1.09) | 1.09 | (0.87, 1.37) | 0.99 | (0.78, 1.26) | 1.06 | (0.78, 1.44) | 1.01 | (0.78, 1.31) | 1.12 | (0.74, 1.71) | 1.10 | (0.85, 1.41) |
| Test of interactionc | 12.50 (1); P=0.0004 | 0.37 (1); P=0.5439 | 0.30 (3); P=0.9598 | |||||||||||||
| Personal crime | ||||||||||||||||
| 1st quartile | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent |
| 2nd quartile | 0.83 | (0.46, 1.50) | 0.52 | (0.33, 0.80) | 0.51 | (0.31, 0.83) | 0.72 | (0.39, 1.33) | 0.71 | (0.45, 1.12) | 0.29 | (0.09, 0.99) | 0.26 | (0.10, 0.67) | 0.65 | (0.25, 1.65) |
| 3rd quartile | 0.60 | (0.27, 1.37) | 0.75 | (0.45, 1.23) | 0.59 | (0.33, 1.07) | 0.87 | (0.42, 1.81) | 0.77 | (0.35, 1.68) | 0.40 | (0.11, 1.54) | 0.30 | (0.09, 1.02) | 0.91 | (0.39, 2.11) |
| 4th quartile | 1.06 | (0.51, 2.21) | 0.48 | (0.27, 0.87) | 0.51 | (0.26, 1.03) | 0.71 | (0.36, 1.39) | 0.67 | (0.23, 1.96) | 0.36 | (0.09, 1.36) | 0.30 | (0.07, 1.25) | 0.75 | (0.32, 1.77) |
| Test of interactionc | 9.59 (3); P=0.0224 | 0.94 (3); P=0.8160 | 6.80 (9); P=0.6581 | |||||||||||||
| Property crime | ||||||||||||||||
| 1st quartile | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent |
| 2nd quartile | 0.72 | (0.35, 1.50) | 0.95 | (0.64, 1.40) | 0.69 | (0.43, 1.09) | 1.22 | (0.65, 2.30) | 1.09 | (0.61, 1.97) | 0.41 | (0.17, 1.02) | 0.95 | (0.31, 2.89) | 0.65 | (0.32, 1.34) |
| 3rd quartile | 0.79 | (0.36, 1.71) | 1.20 | (0.78, 1.84) | 0.90 | (0.56, 1.45) | 1.41 | (0.68, 2.93) | 0.94 | (0.50, 1.77) | 0.84 | (0.49, 1.43) | 1.72 | (0.56, 5.27) | 1.50 | (0.78, 2.92) |
| 4th quartile | 0.93 | (0.43, 2.01) | 1.77 | (1.04, 3.03) | 1.57 | (0.91, 2.70) | 1.61 | (0.91, 2.85) | 2.46 | (1.03, 5.90) | 0.71 | (0.31, 1.59) | 2.11 | (0.67, 6.68) | 1.07 | (0.60, 1.91) |
| Test of interactionc | 2.33 (3); P=0.5061 | 2.98 (3); P=0.3940 | 12.77 (9); P=0.1734 | |||||||||||||
| Number of liquor stores within 5 km | ||||||||||||||||
| None | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent |
| 1 − 10 stores | 1.06 | (0.53, 2.11) | 1.11 | (0.69, 1.78) | 1.07 | (0.57, 2.02) | 1.13 | (0.57, 2.27) | 1.04 | (0.60, 1.78) | 4.31 | (1.50, 12.36) | 1.33 | (0.39, 4.56) | 0.65 | (0.20, 2.10) |
| 11 − 20 stores | 0.67 | (0.27, 1.71) | 1.16 | (0.60, 2.25) | 0.95 | (0.45, 2.05) | 1.13 | (0.48, 2.67) | 1.05 | (0.41, 2.65) | 3.97 | (1.11, 14.11) | 0.57 | (0.08, 4.07) | 0.74 | (0.22, 2.53) |
| ≥21 | 0.68 | (0.30, 1.53) | 1.27 | (0.65, 2.48) | 0.89 | (0.41, 1.96) | 1.37 | (0.55, 3.43) | 2.11 | (0.54, 8.24) | 5.70 | (2.01, 16.18) | 0.62 | (0.14, 2.77) | 0.49 | (0.16, 1.49) |
| Test of interactionc | 5.77 (3); P=0.1231 | 1.09 (3); P=0.7789 | 16.29 (9); P=0.0610 | |||||||||||||
| Random effects | ||||||||||||||||
| School-level intercept | 0.57 | (0.25, 1.31) | 0.56 | (0.24, 1.28) | 0.51 | (0.22, 1.21) | ||||||||||
| Individual-level intercept | 4.01 | (2.51, 6.41) | 3.96 | (2.45, 6.41) | 3.86 | (2.40, 6.20) | ||||||||||
Confidence interval (CI); Adjusted odds ratio (aOR).
Controlling for neighbourhood (median age, % Black, population density) and individual factors (age, sex, race/ethnicity, family affluence, family composition).
Subgroup estimates produced from linear combination of coefficients from product terms between neighbourhood characteristic and subgroup.
Chi-square test (degrees of freedom); P-value.
There was weak evidence for heterogeneity in neighbourhood associations by sex and race/ethnicity. Irrespective of age and race/ethnicity, women living in neighbourhoods within the 2nd vs 1st quartile of personal crime were less likely to binge drink [OR = 0.51; 95% CI (0.31, 0.83)], but this effect could not be differentiated from the association among males [χ2 (3) = 0.94; P = 0.8160]. Likewise, irrespective of age and sex, Blacks living in neighbourhoods within the 2nd vs 1st quartile of personal crime had a lower odds of binge drinking [OR = 0.29; 95% CI (0.09, 0.99)] as did those of other race/ethnicity [OR = 0.26; 95% CI (0.10, 0.67)], though race/ethnic differences were not supported overall [χ2 (9) = 6.80; P = 0.6581]. There was suggestive evidence of a race by liquor store interaction [χ2 (9) = 16.29, P = 0.0610] with only Blacks being more likely to binge drink living near liquor stores within 5 km [e.g. 1–10 stores vs none: OR = 4.31; 95% CI (1.50, 12.36)].
Type of off-premises alcohol outlets
Finally, we substituted liquor stores for beer and wine stores, separately, in our models to gauge sensitivity to type of alcohol outlet (Table 4). In contrast to the adjusted liquor store results, living near two or more beer stores [OR = 1.57; 95% CI (1.03, 2.39)] or wine stores [OR = 2.28; 95% CI (1.52, 3.44)] increased the odds of binge drinking. There was weak evidence of age, sex and race/ethnic interactions, but some associations were of interest. Irrespective of sex and race/ethnicity, only those living near two or more beer stores after age 18 were more likely to binge drink [OR = 2.32; 95% CI (1.52, 3.54)] whereas those living near two or more wine stores were more likely to binge drink both before 18 (OR = 1.77) and after (OR = 2.32). Women, but not men, living near more beer stores [e.g. one store: OR = 1.96; 95% CI (1.01, 3.79)] and living near two or more wine stores [OR = 2.67; 95% CI (1.64, 4.33)] were associated with binge drinking irrespective of age and race/ethnicity. A different pattern emerged by race/ethnicity, irrespective of age and sex. Only among Whites was there an association between binge drinking and beer stores [one store: OR = 2.21; 95% CI (1.06, 4.60)] and wine stores (e.g. two or more: OR = 3.31; 95% CI (1.61, 6.80)]. By contrast, Blacks were only more likely to binge drink living near two or more wine stores [OR = 2.25; 95% CI (1.20, 4.22)].
Table 4.
Odds of past 30-day binge drinking associated with the number of beer and wine stores within 5 km by age, sex and race. NEXT Generation Health Study, 2010-2014 (n = 2745)
| Adjusteda |
By ageab |
By sexab |
By race/ethnicityab |
|||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Under 18 |
18 or Older |
Female |
Male |
White |
Black |
Other |
Hispanic |
|||||||||||
| Type of alcohol outlet | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI |
| Model A - beer stores within 5 km | ||||||||||||||||||
| Fixed effects | ||||||||||||||||||
| None | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent |
| 1 store | 1.53 | (0.90, 2.61) | 1.79 | (0.97, 3.31) | 1.38 | (0.75, 2.54) | 1.96 | (1.01, 3.79) | 1.12 | (0.61, 2.06) | 2.21 | (1.06, 4.60) | 1.14 | (0.77, 1.68) | 0.39 | (0.11, 1.33) | 0.90 | (0.43, 1.87) |
| ≥2 stores | 1.57 | (1.03, 2.39) | 1.46 | (0.83, 2.56) | 1.54 | (1.00, 2.36) | 1.73 | (1.08, 2.77) | 1.30 | (0.74, 2.28) | 2.03 | (0.96, 4.32) | 1.59 | (0.77, 3.28) | 0.59 | (0.17, 2.09) | 0.94 | (0.52, 1.69) |
| Test of interactionc | N/A | 0.49 (2); P=0.7810 | 2.74 (2); P=0.2546 | 9.94 (6); P=0.1272 | ||||||||||||||
| Random effects | ||||||||||||||||||
| School-level intercept | 0.53 | (0.23, 1.21) | 0.56 | (0.25, 1.28) | 0.55 | (0.24, 1.27) | 0.50 | (0.22, 1.14) | ||||||||||
| Individual-level intercept | 3.96 | (2.47, 6.35) | 4.03 | (2.52, 6.46) | 3.98 | (2.47, 6.41) | 3.91 | (2.44, 6.28) | ||||||||||
| Model B - wine stores within 5 km | ||||||||||||||||||
| Fixed effects | ||||||||||||||||||
| None | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent |
| 1 store | 1.58 | (0.99, 2.53) | 1.17 | (0.63, 2.17) | 1.70 | (1.01, 2.85) | 1.76 | (0.93, 3.35) | 1.42 | (0.93, 2.17) | 2.04 | (1.04, 4.03) | 1.48 | (0.72, 3.05) | 0.28 | (0.08, 1.01) | 1.39 | (0.76, 2.52) |
| ≥2 stores | 2.28 | (1.52, 3.44) | 1.77 | (1.05, 3.00) | 2.32 | (1.52, 3.54) | 2.67 | (1.64, 4.33) | 1.75 | (1.00, 3.06) | 3.31 | (1.61, 6.80) | 2.25 | (1.20, 4.22) | 1.31 | (0.49, 3.49) | 1.37 | (0.83, 2.28) |
| Test of interactionc | N/A | 1.59 (2); P=0.4519 | 1.73 (2); P=0.4217 | 10.51 (6); P=0.1047 | ||||||||||||||
| Random effects | ||||||||||||||||||
| School-level intercept | 0.57 | (0.25, 1.33) | 0.60 | (0.26, 1.39) | 0.60 | (0.25, 1.40) | 0.52 | (0.23, 1.19) | ||||||||||
| Individual-level intercept | 4.02 | (2.50, 6.46) | 4.09 | (2.55, 6.56) | 4.04 | (2.50, 6.52) | 3.95 | (2.46, 6.36) | ||||||||||
CI, confidence interval; OR, odds ratio.
Controlling for neighbourhood (disadvantage, personal and property crime, median age, % Black, population density) and individual factors (age, sex, race/ethnicity, family affluence, family composition).
Subgroup estimates produced from linear combination of coefficients from product terms between neighbourhood characteristic and subgroup.
Chi-square test (degrees of freedom); P-value.
Discussion
We examined how neighbourhood disadvantage, crime and the number of liquor stores were associated with individual binge drinking from late adolescence into early adulthood. Then, we examined whether neighbourhood associations differed by age, sex and race/ethnicity, and whether findings varied for beer or wine stores. Consistent with theories that suggest disadvantage and social disorder promote adolescent deviant behaviours,26 we found that neighbourhood disadvantage was associated with binge drinking before adulthood, but not after. As alcohol availability and social acceptance of alcohol increases with age, binge drinking may diminish as a marker of deviance.27 Alternatively, neighbourhood exposure during high school is likely more stable than in early adulthood when environments change (e.g. attending college). Discordance between early adults’ actual environment (e.g. campus) and the surrounding socio-economic environment could have weakened associations. Nonetheless, our findings are consistent with evidence showing that childhood and early adolescent disadvantage is related to drinking20,21 and extend these findings into early adulthood.
That personal crime was associated with less binge drinking was unexpected and seems to contradict social disorganization theory. A prior study noted that witnessing community violence increased alcohol use but being victimized decreased it.28 Even if not witnessing violence, those living in areas of higher violent crime might avoid social situations (e.g. parties) in which alcohol is present or moderate their drinking to prevent victimization. This explanation is consistent with our finding that the association was limited to early adulthood when such social situations are likely to occur. Although our findings of differences by sex or race/ethnicity were inconclusive, this explanation is also consistent with studies showing women and Black youth are especially vulnerable to victimization related to alcohol use and the racial segregation in areas of concentrated disadvantage.29–31
For property crime, we found weak evidence of an overall association with binge drinking or neighbourhood by individual interactions. Nonetheless, there were associations that emerged within certain demographic groups. High property crime was associated with binge drinking only in adulthood and separately among Whites. Prior evidence points to correlations among drinking rates, alcohol outlets and property crime at the ecological level,32,33 and drinking being associated with property crime offending at the individual level,34,35 but our findings may be novel in suggesting that neighbourhood property crime could be related to individual drinking.
With respect to the number of liquor stores within 5 km, we did not find an overall association with binge drinking, but other findings were of interest when we explored differences by individual characteristics and type of alcohol outlet. Blacks were more likely to binge drink living near liquor stores, while Whites were more likely to binge drink living near beer and wine stores. Blacks are more likely than Whites to drink higher alcohol-content beverages, including spirits and malt liquor, and less likely to drink wine.36–38 Although there could be cultural differences in preferences for beverage type, liquor stores are disproportionately concentrated in Black neighbourhoods.39 Associations for beer/wine stores could also be driven by age and sex differences. High school-aged men tend to prefer beer and liquor while women tend to prefer wine, malt beverages and wine coolers.40 Outlet density may be more relevant to early adult drinking when youth can more easily pass off as legally aged and obtain false identification. Binge drinkers tend to prefer beer, whereas preference for wine tends to increase towards mid-adulthood.41,42
One potential study limitation was that we could not account for all factors that might explain both neighbourhood exposures and drinking. For example, parental alcohol history might explain both the type of neighbourhoods their children live in and participants’ drinking habits. Another possible limitation was that our sample is underrepresented by those who skipped or dropped out of school, who may also be more likely to binge drink or come from disadvantaged neighbourhoods. Even among those sampled, these factors could also influence participation in subsequent follow-up waves. This could limit the generalizability of our findings.
In summary, we found that neighbourhood disadvantage was associated with binge drinking in late adolescence, but not early adulthood, and personal crime was inversely associated with it, but mainly in early adulthood. Differences in neighbourhood associations by sex and race/ethnicity were informative, but inconclusive. That liquor stores were associated with binge drinking among Blacks but beer/wine stores were associated with it among Whites is a novel finding that deserves closer follow-up in larger samples. If replicated, communities using zoning and licensing laws to regulate the location and density of alcohol outlets may also need to consider the type of alcohol sold and how outlets are distributed.43 Further, it remains unclear whether neighbourhood contexts have similar relationships to other substances or important mechanisms linking neighbourhood and drinking. For example, peer drinking and norms might mediate these relationships.27,44,45 In terms of interventions, relocating families to less disadvantaged neighbourhoods has shown mixed evidence of benefits.46–48 Policies that increase access to services and employment opportunities, reduce crime, strengthen minimum legal drinking age 21-laws, and raise alcohol taxes49 could be evaluated as part of a multi-faceted public health effort with broader impacts on risky drinking and healthy adolescent development.
Funding
This work was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) [contract number HHSN275201200001I], the National Heart, Lung and Blood Institute (NHLBI), the National Institute on Alcohol Abuse and Alcoholism (NIAAA), and Maternal and Child Health Bureau (MCHB) of the Health Resources and Services Administration (HRSA), with supplemental support from the National Institute on Drug Abuse (NIDA).
References
- 1. Elliott DS, Wilson WJ, Huizinga D, Sampson RJ, Elliott A, Rankin B.. The effects of neighborhood disadvantage on adolescent development. J Res Crime Delinq 1996;33:389–426. [Google Scholar]
- 2. Leventhal T, Brooks-Gunn J.. The neighborhoods they live in: The effects of neighborhood residence on child and adolescent outcomes. Psychol Bull 2000;126:309–37. [DOI] [PubMed] [Google Scholar]
- 3. Ennett ST, Flewelling RL, Lindrooth RC, Norton EC.. School and neighborhood characteristics associated with school rates of alcohol, cigarette, and marijuana use. J Health Soc Behav 1997;38:55.. [PubMed] [Google Scholar]
- 4. Miller JW, Naimi TS, Brewer RD, Jones SE.. Binge drinking and associated health risk behaviors among high school students. Pediatrics 2007;119:76–85. [DOI] [PubMed] [Google Scholar]
- 5. Kuntsche E, Kuntsche S, Thrul J, Gmel G.. Binge drinking: health impact, prevalence, correlates and interventions. Psychol Heal 2017;32:976–1017. [DOI] [PubMed] [Google Scholar]
- 6. Kubrin CE, Weitzer R.. New directions in social disorganization theory. J Res Crime Delinq 2003;40:374–402. [Google Scholar]
- 7. Zucker RA. Anticipating problem alcohol use developmentally from childhood into middle adulthood: what have we learned? Addiction 2008;103:100–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Boardman JD, Finch BK, Ellison CG, Williams DR, Jackson JS.. Neighborhood disadvantage, stress, and drug use among adults. J Health Soc Behav 2001;42:151–65. [PubMed] [Google Scholar]
- 9. Fagan AA, Wright EM, Pinchevsky GM.. Exposure to violence, substance use, and neighborhood context. Soc Sci Res 2015;49:314–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Jencks C, Mayer SE.. The social consequences of growing up in a poor neighborhood In: Lynn LE, McGeary MFH (eds). Inner-City Poverty in the United States. Washington: National Academy Press, 1990, pp. 111–86. [Google Scholar]
- 11. Foster S, Trapp G, Hooper P, Oddy WH, Wood L, Knuiman M.. Liquor landscapes: Does access to alcohol outlets influence alcohol consumption in young adults? Heal Place 2017;45:17–23. [DOI] [PubMed] [Google Scholar]
- 12. Livingston M, Chikritzhs T, Room R.. Changing the density of alcohol outlets to reduce alcohol-related problems. Drug Alcohol Rev 2007;26:557–66. [DOI] [PubMed] [Google Scholar]
- 13. Gruenewald PJ, Freisthler B, Remer L, LaScala EA, Treno A.. Ecological models of alcohol outlets and violent assaults: Crime potentials and geospatial analysis. Addiction 2006;101:666–77. [DOI] [PubMed] [Google Scholar]
- 14. Pollack CE, Cubbin C, Ahn D, Winkleby M.. Neighbourhood deprivation and alcohol consumption: Does the availability of alcohol play a role? Int J Epidemiol 2005;34:772–80. [DOI] [PubMed] [Google Scholar]
- 15. Gmel G, Holmes J, Studer J.. Are alcohol outlet densities strongly associated with alcohol-related outcomes? A critical review of recent evidence. Drug Alcohol Rev 2016;35:40–54. [DOI] [PubMed] [Google Scholar]
- 16. Karriker-Jaffe KJ. Areas of disadvantage: A systematic review of effects of area-level socioeconomic status on substance use outcomes. Drug Alcohol Rev 2011;30:84–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Bryden A, Roberts B, Petticrew M, McKee M.. A systematic review of the influence of community level social factors on alcohol use. Health Place 2013;21:70–85. [DOI] [PubMed] [Google Scholar]
- 18. Jackson N, Denny S, Ameratunga S.. Social and socio-demographic neighborhood effects on adolescent alcohol use: A systematic review of multi-level studies. Soc Sci Med 2014;115:10–20. [DOI] [PubMed] [Google Scholar]
- 19. Bryden A, Roberts B, McKee M, Petticrew M.. A systematic review of the influence on alcohol use of community level availability and marketing of alcohol. Health Place 2012;18:349–57. [DOI] [PubMed] [Google Scholar]
- 20. Cambron C, Kosterman R, Catalano RF, Guttmannova K, Hawkins JD.. Neighborhood, family, and peer factors associated with early adolescent smoking and alcohol use. J Youth Adolesc 2018;47:369–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Karriker-Jaffe KJ, Lönn SL, Cook WK, Kendler KS, Sundquist K.. Chains of risk for alcohol use disorder: Mediators of exposure to neighborhood deprivation in early and middle childhood. Health Place 2018;50:16–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Krieg AG, Kuhl DC.. Race, adolescent binge drinking, and the context of neighborhood exposure. Deviant Behav 2016;37:615–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Barr PB. Neighborhood conditions and trajectories of alcohol use and misuse across the early life course. Health Place 2018;51:36–44. [DOI] [PubMed] [Google Scholar]
- 24. Currie C, Molcho M, Boyce W, Holstein B, Torsheim T, Richter M.. Researching health inequalities in adolescents: The development of the Health Behaviour in School-Aged Children (HBSC) Family Affluence Scale. Soc Sci Med 2008;66:1429–36. [DOI] [PubMed] [Google Scholar]
- 25.StataCorp. Stata Statistical Software: Release 14. College Station, TX: StataCorp, LP; 2015.
- 26. Sampson RJ, Morenoff JD, Gannon-Rowley T.. Assessing “Neighborhood Effects”: Social processes and new directions in research. Annu Rev Sociol 2002;28:443–78. [Google Scholar]
- 27. Song E-Y, Smiler AP, Wagoner KG, Wolfson M.. Everyone says it’s ok: Adolescents’ perceptions of peer, parent, and community alcohol norms, alcohol consumption, and alcohol-related consequences. Subst Use Misuse 2012;47:86–98. [DOI] [PubMed] [Google Scholar]
- 28. Mrug S, Windle M.. Initiation of alcohol use in early adolescence: Links with exposure to community violence across time. Addict Behav 2009;34:779–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Devries KM, Child JC, Bacchus LJ, Mak J, Falder G, Graham K.. Intimate partner violence victimization and alcohol consumption in women: A systematic review and meta-analysis. Addiction 2014;109:379–91. [DOI] [PubMed] [Google Scholar]
- 30. Lo CC, Howell RJ, Cheng TC.. Explaining Black–White differences in homicide victimization. Aggress Violent Behav 2013;18:125–34. [Google Scholar]
- 31. Peterson RD, Krivo LJ.. Racial segregation, the concentration of disadvantage, and Black and White Homicide victimization. Sociol Forum 1999;14:465–93. [Google Scholar]
- 32. Martin SE, Maxwell CD, White HR, Zhang Y.. Trends in alcohol use, cocaine use, and crime: 1989–1998. J Drug Issues 2004;34:333–59. [Google Scholar]
- 33. Groff ER, Lockwood B.. Criminogenic facilities and crime across street segments in Philadelphia. J Res Crime Delinq 2014;51:277–314. [Google Scholar]
- 34. Popovici I, Homer JF, Fang H, French MT.. Alcohol use and crime: Findings from a longitudinal sample of U.S. Adolescents and young adults. Alcohol Clin Exp Res 2012;36:532–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Fergusson DM, Horwood LJ.. Alcohol abuse and crime: A fixed-effects regression analysis. Addiction 2000;95:1525–36. [DOI] [PubMed] [Google Scholar]
- 36. Dawson DA. Beyond black, white and Hispanic: Race, ethnic origin and drinking patterns in the United States. J Subst Abuse 1998;10:321–39. [DOI] [PubMed] [Google Scholar]
- 37. Bluthenthal RN, BrownTaylor D, Guzmán-Becerra N, Robinson PL.. Characteristics of malt liquor beer drinkers in a low-income, racial minority community sample. Alcohol Clin Exp Res 2005;29:402–9. [DOI] [PubMed] [Google Scholar]
- 38. Kerr WC, Patterson D, Greenfield TK.. Differences in the measured alcohol content of drinks between Black, White and Hispanic men and women in a US national sample. Addiction 2009;104:1503–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. LaVeist TA, Wallace JM.. Health risk and inequitable distribution of liquor stores in African American neighborhood. Soc Sci Med 2000;51:613–7. [DOI] [PubMed] [Google Scholar]
- 40. Siegel MB, Naimi TS, Cremeens JL, Nelson DE.. Alcoholic beverage preferences and associated drinking patterns and risk behaviors among high school youth. Am J Prev Med 2011;40:419–26. [DOI] [PubMed] [Google Scholar]
- 41. Naimi TS, Brewer RD, Miller JW, Okoro C, Mehrotra C.. What do binge drinkers drink? Implications for alcohol control policy. Am J Prev Med 2007;33:188–93. [DOI] [PubMed] [Google Scholar]
- 42. Klatsky AL, Armstrong MA, Kipp H.. Correlates of alcoholic beverage preference: Traits of persons who choose wine, liquor or beer. Br J Addict 1990;85:1279–89. [DOI] [PubMed] [Google Scholar]
- 43. Campbell CA, Hahn RA, Elder R. et al. The effectiveness of limiting alcohol outlet density as a means of reducing excessive alcohol consumption and alcohol-related harms. Am J Prev Med 2009;37:556–69. [DOI] [PubMed] [Google Scholar]
- 44. Martino SC, Ellickson PL, McCaffrey DF.. Multiple trajectories of peer and parental influence and their association with the development of adolescent heavy drinking. Addict Behav 2009;34:693–700. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Rudolph KE, Sofrygin O, Schmidt NM. et al. Mediation of neighborhood effects on adolescent substance use by the school and peer environments. Epidemiology 2018;29:590–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Linton SL, Cooper HLF, Luo R. et al. People and places: Relocating to neighborhoods with better economic and social conditions is associated with less risky drug/alcohol network characteristics among African American adults in Atlanta, GA. Drug Alcohol Depend 2016;160:30–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Fauth RC, Leventhal T, Brooks-Gunn J.. Short-term effects of moving from public housing in poor to middle-class neighborhoods on low-income, minority adults’ outcomes. Soc Sci Med 2004;59:2271–84. [DOI] [PubMed] [Google Scholar]
- 48. Fauth RC, Leventhal T, Brooks-Gunn J.. Welcome to the neighborhood? Long-term impacts of moving to low-poverty neighborhoods on poor children’s and adolescents’ outcomes. J Res Adolesc 2007;17:249–84. [Google Scholar]
- 49. Fairman BJ, Simons-Morton BG, Haynie DL. et al. State alcohol policies, taxes, and availability as predictors of adolescent binge drinking trajectories into early adulthood. Addiction. 2019;114:1173–82. [DOI] [PMC free article] [PubMed] [Google Scholar]