Abstract
Objectives
To investigate sociodemographic disparities in alcohol environments and their relationship with adolescent drinking.
Methods
ArcMap is used to geocode and map alcohol license data to constructed circular buffers centered at 14,595 residences of households with children in the California Health Interview Survey. Commercial sources of alcohol in each buffer are calculated. Multivariate logistic regression is used to discern effects of alcohol sales on adolescents’ drinking from their individual, family, and neighborhood characteristics.
Results
Alcohol availability, measured by mean and median number of license, is significantly higher around residences of minority and lower-income families. Binge drinking and driving after drinking among adolescents ages 12−17 are significantly associated with alcohol retailers within 0.5 mile from home. Counterfactual simulation shows that if alcohol sales is cut from mean number of alcohol outlets around the lowest-income quartile households to that of the highest quartile, prevalence of binge drinking would fall from 6.4% to 5.6%, and driving after drinking from 7.9% to 5.9%.
Conclusions
Alcohol outlets are concentrated in disadvantaged neighborhoods and can contribute to youth's drinking. Sociodemographic disparities in health risks can be worsened by alcohol-related problems.
INTRODUCTION
Despite all federal, state, and local interventions, underage drinking continues to be a serious problem. Among adolescents, 17.6% used alcohol in the past 30 days, 11.1% were binge drinkers, and 2.7% were heavy drinkers.1 Health and social problems associated with youth drinking include motor vehicle crashes,2,3 violence,4 risky sexual behaviors,5,6 assault and rapes,7 and brain impairment.8−11 and cause substantial societal costs.12 Drinking at an earlier age also increases the risk of addiction and other alcohol-related problems in adulthood.13−15 In 2007, the Surgeon General responded to this problem in the “Call to Action to Prevent and Reduce Underage Drinking”, which also emphasized the influence of environments .16
Underage drinkers obtain their alcoholic beverages from a variety of sources, including parents’ stocks, friends, parties, and beginning in teenage years, from commercial outlets.17 Pseudo-underage buyers purchased alcohol with high success rates from both on-premise and off-premise establishments.18,19 Sales to minors were significantly and positively related to the percentage of Hispanic residents and higher neighborhood population density.20
As long as adolescents can obtain alcohol from commercial sources, neighborhood outlets are likely to play a role in underage drinking. Rhee et al. argue that environments play a greater role in drinking initiation, whereas genetics are more important for alcohol dependence.21 Perceived alcohol availability was significantly associated with higher levels of alcohol consumption for males and drinking in public locations for females,22,23 and city-level outlet density measures were associated with both youth's drinking and driving, and riding with a drinking driver.24
Differential alcohol environments may exacerbate health disparities across sociodemographic groups. LaVeist and Wallace found more liquor stores per capita in low-income and predominantly black census tracts in Baltimore.25 Gorman and Speer found retail liquor outlets abundantly located in poor and minority neighborhoods within one city in New Jersey.26 Only one published national study exists and it reported higher densities of liquor stores in zip codes of higher percentages of Blacks and lower-income non-Whites.27 While that study covered all urban areas in the United States, urban zip codes with a mean land area of 40.1 square miles and mean population of 21,920 persons27 are arguably too large to represent neighborhoods. Even census tracts can range from 0.04 square miles to 322 square miles just in Los Angeles County.
The two objectives of this study are: 1) describe the quantity and locational pattern of alcohol retailers in small areas around individual homes (rather than using administrative definitions like census tracts); 2) examine relationships between alcohol environments and adolescent drinking. Our data covers the whole state of California while the methodology focuses on individuals’ spatial accessibility to alcohol sales.
METHODS
Data
Data on alcohol outlets come from the California Department of Alcoholic Beverage Control (ABC) database and include address and license types of all alcohol retailers in the state.28 We classify license types into off-sale (for consumption elsewhere, like liquor stores) and on-sale (for consumption on premises, like restaurants). In 2003, California had 30,650 active on-sale licenses and 21,836 active off-sale licenses. Details are available at http://www.abc.ca.gov/.
The California Health Interview Survey (CHIS) is a computer-assisted telephone interview, using a two-stage, geographically stratified random-digit-dial design that attempts to interview one adult, one adolescent, and gets information on one child in households with children. The survey is representative of the state's non-institutionalized population living in households. Details are available at http://www.chis.ucla.edu/.
The CHIS 2003 includes survey data for 42,044 adults, 4,010 adolescents, and 8,526 children, who are linked by family identifiers. We excluded 3,679 households in rural areas because their environments are incomparable. For the analysis of alcohol environments, we focus on 14,595 households with children age under 18 (not all of these households had complete youth survey). For the analysis of adolescent drinking, we analyzed 3,660 adolescents ages 12−17. A sub-sample of 687 adolescents ages 16−17 who ever had a few sips of alcoholic drinks was used for the analysis of adolescent driving after drinking.
Measures of Alcohol Environments
Our definition of alcohol environments is based on distance from homes. We draw circles with radii of 0.1 mile, 0.5 mile, 1.0 mile and 2.0 mile centered at the respondents’ residence using ArcMap version 9.1. We first look at immediate distances using 0.1 mile radius circles and circular bands between 0.1 mile and 0.5 mile radii. Outlets within these distances may be most problematic because of their close proximity to residences. A distance of 0.5 mile is about a 10 minute walk29 and within the reach of adolescents. Outlets beyond easy walking distance are examined in circular bands between 0.5 and 1.0 mile radii and between 1.0 and 2.0 mile radii (and all 4 constructed buffers are mutually exclusive). We map the business locations in ABC data to the buffers around each household and calculate the number of alcohol retailers within each buffer.
Previous research focused on density measures such as the number of establishments per city, per residents, or per roadway miles.27,30−32 We do not normalize outlet measures in such ways. Conceptually, the raw count in each buffer is preferable to outlet density measures in a predefined geographic area like census tracts because a person may live close to outlets, yet the density of outlets may be minimally small if the area includes large parts of deserts or mountains. Similarly, in densely populated urban areas, the population definition may show low densities even when most residents live within walking distance of alcohol outlets.
Statistical Methods
Sociodemographic Disparities in Alcohol Environments
We compare the mean and median number of alcohol outlets (all licenses, on-sale, and off-sale, respectively) across race/ethnicity groups (non-Hispanic White, non-Hispanic Black, Hispanic, Asian/Pacific Islander, and other races) and income groups (incomes quartiles derived from self-reported total household annual income before tax). We then stratify by both race/ethnicity and income. We also use a zero-inflated Poisson regression with number of outlets as the dependent variable, race/ethnicity and income as the key explanatory variables, and control for population density in the census tracts. We estimate this model separately for each definition of the dependent variable (all licenses, on-sale, and off-sale) within each buffer. The data includes all households with children age under 18.
Alcohol Sales and Adolescent Drinking
We analyze three dichotomous dependent variables for adolescent drinking using logistic regression: at least one alcoholic drink in the past 30 days, at least one heavy drinking episode (5 drinks in a row, also referred to as binge drinking) in the past 30 days, and ever driving after drinking. The primary explanatory variables are the number of alcohol outlets within the 0.5 mile radii, 0.5 mile −1.0 mile bands, and 1.0 mile-2.0 mile bands. For each dependent variable, we estimate two models that differ by the key explanatory variables. The first model has total number of licenses as the key explanatory variable. The second model has off-sale stores and on-sale establishments as key explanatory variables. The latter model is meant to determine what type of outlets has predictive power of adolescent drinking, because the underlying processes for youth to illegally obtain alcoholic beverages may differ.
Additional explanatory variables included in all models are adolescents’ characteristics (gender, age, race, work for pay in the past 12 months, current smoker, marijuana use in the past 30 days), family characteristics (household income, parents’ marital status), parent drinking behavior (indicators whether parent/guardian self-reported any heavy drinking episode – defined as 5 drinks in a row in the past 30 days, and excess drinking – defined as consumed more than 60 drinks per month), and neighborhood sociodemographics (census tract total population, tract median household income, percent of White population, and percent of Black population; data extracted from US Census 2000). All regression models use robust standard errors to account for clustering data due to CHIS's multi-stage sample design. First, the state was divided into 44 geographic sampling strata, including 41 single-county strata and three multi-county strata comprised of the 17 remaining counties in California. Second, within each geographic stratum, residential telephone numbers were selected through random-digit dial sampling. The regression is also weighted to control for differential sampling rates within geographic stratum and race/ethnicity groups.
To improve the interpretation of logistic regression coefficients, we use a counterfactual simulation by changing levels of alcohol availability in adolescents’ neighborhoods and predicting the resulting prevalence of adolescent drinking using the estimated model. We only change the key explanatory, keeping all other variables the same. This provides the adjusted difference in the prevalence of a drinking measure between two levels of alcohol availability, i.e. it accounts for all individual, family, and neighborhood sociodemographic characteristics in the model except the alcohol environments. For the differences in alcohol environments, we contrast the average number of outlets around Asian vs White households, or low vs high income households.
RESULTS
Disparities in Alcohol Environments
Table 1 provides descriptive statistics of the sample, which is divided into four quartiles based on gross annual household income: less than $24,000, between $24,000 and $50,000, between $50,000 and $90,000, and more than $90,000. Fewer than 11% of non-Hispanic Whites belong to the bottom income quartile, compared to 32.0% of non-Hispanic Blacks, 50.4% of Hispanic, 20.8% of Asian/Pacific Islanders, and 32.9% of other races. In contrast, 36.0% of non-Hispanic Whites, 15.1% of non-Hispanic Blacks, 4.7% of Hispanic, 29.1% of Asian/Pacific Islander, and 13.2% of other races are in the top income quartile.
Table 1.
Descriptive Statistics for CHIS 2003
| Variable | Mean | Std. Dev. |
|---|---|---|
| Adult sample (Households with children under 18, n1=14,595) | ||
| Non-Hispanic Whitea (47.6%) | ||
| Lowest income quartile | 0.108 | (0.31) |
| 2nd income quartile | 0.225 | (0.42) |
| 3rd income quartile | 0.307 | (0.46) |
| Highest income quartile | 0.360 | (0.48) |
| Non-Hispanic Blacka (7.1%) | ||
| Lowest income quartile | 0.320 | (0.47) |
| 2nd income quartile | 0.277 | (0.45) |
| 3rd income quartile | 0.252 | (0.43) |
| Highest income quartile | 0.151 | (0.36) |
| Hispanica (30.1%) | ||
| Lowest income quartile | 0.504 | (0.50) |
| 2nd income quartile | 0.330 | (0.47) |
| 3rd income quartile | 0.119 | (0.32) |
| Highest income quartile | 0.047 | (0.21) |
| Asian & Pacific Islandera (11.4%) | ||
| Lowest income quartile | 0.208 | (0.41) |
| 2nd income quartile | 0.250 | (0.43) |
| 3rd income quartile | 0.251 | (0.43) |
| Highest income quartile | 0.291 | (0.45) |
| Other racesa (3.8%) | ||
| Lowest income quartile ( ≤ $24,000) | 0.329 | (0.47) |
| 2nd quartile (between $24,000 & $50,000) | 0.316 | (0.47) |
| 3rd quartile (between $50,000 & $90,000) | 0.224 | (0.42) |
| Highest income quartile (> $90,000) | 0.132 | (0.34) |
| Adolescent sample (n2=3,660) | ||
| Ever had acoholic drinks | 0.351 | (0.48) |
| At least 1 drink past 30 days | 0.150 | (0.36) |
| Any binge drinking past 30 days | 0.056 | (0.23) |
| Ever driving after drinkingb | 0.060 | (0.24) |
| Current smoker | 0.050 | (0.22) |
| Marijuana use past 30 daysc | 0.050 | (0.22) |
| Female | 0.489 | (0.50) |
| Age | 14.362 | (1.67) |
| Teens worked for pay past 12 months | 0.405 | (0.49) |
| Parents married or living with a partner | 0.822 | (0.38) |
| Parent's excess drinking | 0.021 | (0.13) |
| Parent's binge drinking past month | 0.131 | (0.34) |
Note:
Race/ethnicity is of the interviewed adult.
Denominator is adolescents aged 16 or older who ever had more than a few sips of alcoholic drinks.
Denominator is adolescents with parent/guardian's permission to ask questions about illicit drug use (98.7%).
All statistics are weighted.
Average age in the adolescent sample is 14.3 years, reflecting the period of drinking initiation. However, the survey did not ask for age when drinking the first time. About 35% of them reported ever having more than just a few sips of alcoholic drinks. Fifteen percent reported having at least 1 drink, and 5.6% having at least one heavy drinking episode in the past 30 days. Five percent reported themselves as current smoker (at least 1 cigarette a day in the past 30 days), and 5.0% reported marijuana use in the past 30 days. Of those aged 16 or 17 who ever consumed alcohol, 6.0% reported ever driving after drinking.
Stratified by income and race/ethnicity, Table 2 shows the mean number of alcohol outlets within different buffers. Panel A shows that compared to non-Hispanic Whites, people of other race groups are surrounded by more alcohol outlets regardless of the size of the buffers. For instance, within 0.1 mile, there is on average 0.21 outlets around White residences compared to 0.24, 0.39, and 0.33 outlets around residences of Black, Hispanic and Asian respondents (differences are statistically significant at 0.1% level). Panel B shows people in lower income quartiles are surrounded by more alcohol outlets. This locational pattern is observed even within each race group. Panels C, D, E, F, and G show the same distribution pattern across income groups within each race. We obtained consistently the same results in the sensitivity analyses: comparing the median number of outlets; separation of off-sales from on-sales; and zero-inflated Poisson regression model with income and race/ethnicity as key predictors of alcohol outlets.
Table 2.
Mean Number of All Alcohol Outlets Around Residences by Race/Ethnicity and Incomea
| 0.1 mile radii | 0.1−0.5 mile bands | 0.5−1.0 mile bands | 1.0−2.0 mile bands | |||
|---|---|---|---|---|---|---|
|
All Race/Ethnicity Groups |
0.30 |
6.97 |
19.20 |
61.55 |
||
| A |
All Income Groups |
White (Ref. Grp) | 0.21 | 5.27 | 15.23 | 49.78 |
| Black | 0.24 *** | 6.22 *** | 17.50 *** | 63.63 *** | ||
| Hispanic | 0.39 *** | 8.10 *** | 21.79 *** | 68.38 *** | ||
| Asian/Pacific Islander | 0.33 *** | 9.18 *** | 24.04 *** | 74.53 *** | ||
| Other races |
0.36 *** |
6.22 *** |
19.15 ** |
62.44 ** |
||
| B |
All Race/Ethnicity Groups |
Lowest Income Quartile (Ref. Grp) | 0.44 | 9.09 | 23.70 | 74.55 |
| 2nd Income Quartile | 0.34 *** | 7.03 *** | 19.37 *** | 59.81 *** | ||
| 3rd Income Quartile | 0.20 *** | 5.66 *** | 16.16 *** | 52.52 *** | ||
| Highest Income Quartile |
0.16 *** |
5.21 *** |
15.67 *** |
54.54 *** |
||
| C |
White |
Lowest Income Quartile (Ref. Grp) | 0.29 | 6.24 | 16.87 | 49.66 |
| 2nd Income Quartile | 0.28 | 5.78 * | 15.34 | 48.09 | ||
| 3rd Income Quartile | 0.15 *** | 4.67 *** | 14.14 ** | 45.47 | ||
| Highest Income Quartile |
0.18 ** |
5.19 ** |
15.60 |
44.54 * |
||
| D |
Black |
Lowest Income Quartile (Ref. Grp) | 0.30 | 8.41 | 21.46 | 78.42 |
| 2nd Income Quartile | 0.24 * | 5.43 *** | 16.84 | 60.48 * | ||
| 3rd Income Quartile | 0.24 * | 5.46 *** | 16.17 * | 57.63 ** | ||
| Highest Income Quartile |
0.10 ** |
4.29 *** |
12.58 ** |
48.16 *** |
||
| E |
Hispanic |
Lowest Income Quartile (Ref. Grp) | 0.47 | 9.38 | 24.72 | 76.79 |
| 2nd Income Quartile | 0.35 *** | 7.28 *** | 19.88 *** | 62.48 *** | ||
| 3rd Income Quartile | 0.23 *** | 6.00 *** | 16.93 *** | 54.90 *** | ||
| Highest Income Quartile |
0.15 *** |
5.43 *** |
16.00 *** |
53.93 *** |
||
| F |
Asian/Pacific Islander |
Lowest Income Quartile (Ref. Grp) | 0.53 | 12.02 | 29.54 | 96.67 |
| 2nd Income Quartile | 0.49 | 11.01 * | 29.83 | 78.97 ** | ||
| 3rd Income Quartile | 0.21 ** | 9.01 ** | 22.10 ** | 70.54 *** | ||
| Highest Income Quartile |
0.14 *** |
5.72 *** |
16.80 *** |
58.31 *** |
||
| G | Other races | Lowest Income Quartile (Ref. Grp) | 0.44 | 8.83 | 21.20 | 68.27 |
| 2nd Income Quartile | 0.34 | 5.24 ** | 19.69 | 65.23 | ||
| 3rd Income Quartile | 0.45 | 5.80 ** | 18.44 | 62.48 | ||
| Highest Income Quartile | 0.06 ** | 2.79 *** | 13.98 ** | 41.21 * |
Note:
Sample includes 14,595 households with children ages 0−17.
Band 0.1 mile-0.5 mile: the area between the 0.1 mile radius and 0.5 mile radius.
Band 0.5 mile-1.0 mile: the area between the 0.5 mile radius and 1.0 mile radius.
Band 1.0 mile-2.0 mile: the area between the 1.0 mile radius and 2.0 mile radius.
Household income cutoff points: 1st quartile ≤ $24,000; $24,000 <2nd quartile ≤ $50,000; $50,000 < 3rd quartile ≤ $90,000; $90,000 <4th quartile.
indicates significance at 5% level
significance at the 1% level
significance at 0.1% level.
Reference group is in Italic in each panel. All statistics are weighted.
Alcohol Sales and Adolescent Drinking
The results from 6 logistic regression models (3 dependent variables × 2 model specifications) are reported in Table 3. In Model 1 we find that total number of alcohol outlets within 0.5 mile from homes are significantly associated with adolescent binge drinking (p < 0.001) and driving after drinking (p < 0.001), after taking into account adolescents’ individual and family characteristics, parent/guardian's drinking behavior, and neighborhood sociodemographics. Alcohol outlets located further away appear to have no relationship with any measures of adolescent drinking.
Table 3.
Effects of Alcohol Outlets on Adolescent Drinking
| (A) | (B) | (C) | |||||
|---|---|---|---|---|---|---|---|
| Explanatory Variable | 1 drink past 30 days | 5 drinks past 30 days | Ever Driving after Driking | ||||
| OR | 95% C.I. | OR | 95% C.I. | OR | 95% C.I. | ||
|
Model 1 |
All licenses 0.5 mile radius | 1.01 | (0.90, 1.03) | 1.03 | (1.01, 1.05) *** | 1.11 | (1.05, 1.17) *** |
| All licenses 0.5−1.0 mile band | 0.99 | (0.98, 1.01) | 0.98 | (0.97, 1.01) | 0.96 | (0.90, 1.01) | |
| All licenses 1.0−2.0 mile band |
1.00 |
(1.00, 1.00) |
1.00 |
(0.99, 1.01) |
1.00 |
(0.99, 1.01) |
|
| Off-sales 0.5 mile radius | 1.00 | (0.94, 1.07) | 1.03 | (1.01, 1.07) * | 1.06 | (0.87, 1.30) | |
| Off-sales 0.5−1.0 mile band | 0.98 | (0.94, 1.02) | 0.99 | (0.97, 1.01) | 0.90 | (0.75, 1.07) | |
| Off-sales 1.0−2.0 mile band | 1.00 | (0.98, 1.01) | 1.00 | (0.98, 1.01) | 1.05 | (0.99, 1.12) | |
| On-sales 0.5 mile radius | 1.01 | (0.99, 1.03) | 1.03 | (1.01, 1.07) * | 1.14 | (1.05, 1.23) *** | |
| On-sales 0.5−1.0 mile band | 1.00 | (0.98, 1.02) | 0.99 | (0.97, 1.01) | 0.99 | (0.88, 1.11) | |
| On-sales 1.0−2.0 mile band |
1.00 |
(1.00, 1.01) |
1.00 |
(1.00, 1.03) |
0.97 |
(0.93, 1.01) |
|
|
Model 2 |
Female | 0.95 | (0.73, 1.25) | 0.65 | (0.42, 1.00) * | 0.57 | (0.23, 1.44) |
| Age | 1.60 | (1.46, 1.76) *** | 1.79 | (1.58, 2.03) *** | 2.05 | (0.77, 5.48) | |
| Hispanic | 1.15 | (0.78, 1.67) | 1.17 | (0.57, 2.39) | 4.31 | (1.44, 12.93) ** | |
| Asian/Pacific Islander | 0.56 | (0.31, 1.01) * | 0.59 | (0.24, 1.40) | 1.05 | (0.08, 13.91) | |
| Black | 0.42 | (0.19, 0.90) * | 0.56 | (0.14, 2.20) | 1.21 | (0.16, 9.16) | |
| Other races | 0.82 | (0.41, 1.63) | 1.07 | (0.34, 3.34) | 2.07 | (0.36, 11.97) | |
| Work for pay past 12 months | 1.14 | (0.85, 1.52) | 1.18 | (0.74, 1.87) | 1.00 | (0.38, 2.66) | |
| Current smoker | 2.63 | (1.48, 4.68) *** | 4.30 | (2.15, 8.60) *** | 7.93 | (2.93, 21.45) *** | |
| Marijuana use past 30 days |
15.66 |
(9.03, 27.15) *** |
17.91 |
(9.59, 33.44) *** |
5.42 |
(2.44, 12.04) *** |
|
| 2nd income quartile | 1.39 | (0.91, 2.13) | 0.86 | (0.44, 1.68) | 6.89 | (1.71, 27.75) ** | |
| 3rd income quartile | 1.17 | (0.72, 1.90) | 0.78 | (0.38, 1.60) | 4.69 | (1.19, 18.55) * | |
| Highest income quartile | 1.46 | (0.85, 2.48) | 1.17 | (0.56, 2.43) | 11.20 | (2.12, 59.16) ** | |
| Parents married or living w/ a partner | 0.59 | (0.42, 0.85) ** | 0.65 | (0.36, 1.16) | 0.47 | (0.17, 1.32) | |
| Parent's excess drinking | 0.28 | (0.08, 1.00) | 0.14 | (0.01, 1.38) | 0.30 | (0.11, 2.22) | |
| Parent's binge drinking |
1.51 |
(0.98, 2.33) |
1.64 |
(0.92, 2.91) |
1.55 |
(0.65, 3.69) |
|
| Census tract population | 1.00 | (1.00, 1.00) | 1.00 | (1.00, 1.00) | 1.00 | (1.00, 1.00) | |
| Tract median hshd income | 0.94 | (0.62, 1.43) | 1.02 | (0.50, 2.08) | 0.83 | (0.23, 2.99) | |
| Tract % White population | 0.61 | (0.28, 1.29) | 0.99 | (0.25, 3.96) | 3.32 | (0.46, 24.12) | |
| Tract % Black population | 0.42 | (0.07, 2.35) | 0.98 | (0.04, 23.29) | 9.09 | (2.46, 34.23) * | |
Note:
Sample in (A) and (B) includes 3,660 adolescents ages 12−17 and sample in (C) includes 687 youth ages 16−17.
Model 1 includes the same other explanatory variables as Model 2 but their estimates are not reported due to space limiation.
Household income: 1st quartile ≤ $24,000; $24,000 <2nd quartile ≤ $50,000; $50,000 < 3rd quartile ≤ $90,000; $90,000 <4th quartile.
White is the reference group for race; Lowest income quartile is the reference group for income.
indicates significance at 5% level
significance at the 1% level
significance at 0.1% level.
All statistics are weighted.
Model 2 splits the total number of outlets into off-sales and on-sales. Both types of outlets located within 0.5 mile are independently and significantly associated with binge drinking and the magnitude of their effects is about the same. On-sale retailers located within 0.5 mile are significantly associated with driving after drinking.
With a binge drinking rate of 5.6% among adolescents, an odds ratio of 1.027 (all outlets within 0.5 mile and binge drinking) corresponds an increase of 0.1 percentage point for a single additional alcohol outlet within 0.5 mile. Table 2 shows that the difference in the mean number of all alcohol outlets located within the 0.5 mile radii between the bottom and top income quartile is about 4.
Using Model 1, we simulate changes in the prevalence of adolescent drinking following changes in the alcohol environment within the 0.5 mile radii. Table 4 shows that if everyone lived in neighborhoods that had the number of alcohol outlets equal to the average Asian households do, the prevalence of adolescent binge drinking and driving after drinking would be 6.4% and 7.9%. If all adolescents were exposed to the number of alcohol outlets that is the average for White households, the corresponding statistics would drop to 5.6% and 6.0%. Table 4 also shows the simulation results corresponding to alcohol environments around the income quartiles and the two groups exposed most and least to alcohol sales.
Table 4.
Simulated Prevalence of Adolescent Drinking by Alcohol Availability
| Mean Number of Alcohol Outlets within 0.5 mile from Residences | Heavy Drinking Episode | Ever Driving After Drinking |
|---|---|---|
| Asian Level (Ua = 9.51) | 0.064 | 0.079 |
| White Level (Uw = 5.48) |
0.056 |
0.060 |
| Lowest Income Level (Ul = 9.53) | 0.064 | 0.079 |
| Highest Income Level (Uh = 5.37) |
0.056 |
0.059 |
| Asian Lowest Income Level (Ual = 12.55) | 0.067 | 0.098 |
| White Highest Income Level (Uwh = 5.37) | 0.056 | 0.059 |
Note:
Regression model 1 with total number of alcohol outlets as key explanatory variables is used.
Only alcohol outlets in 0.5 mile bands changed.
95% confidence intervals are reported in parentheses. All statistics are weighted.
Age, current smoking, and marijuana use are positively and significantly associated with adolescent drinking. Girls are less likely to binge-drink. Asian and Black adolescents are less likely to have any drink. Family income does not predict the first two measures of drinking but is significantly associated with driving after drinking. Parents’ marital status and drinking behavior do not predict youth drinking except that living with married parents is a protective factor for at least one drink. Hispanic adolescents are much more likely to drive after drinking even after accounting for other factors. Percent of blacks in neighborhoods is significantly associated with youth driving after drinking.
DISCUSSION
This paper finds that alcohol outlets within walking distance from homes are associated with two adverse alcohol behaviors among adolescents: increased binge drinking and driving after drinking. The potential effects of differences in alcohol environments could be substantial. Based on our estimated model, changing the number of outlets within 0.5 miles from 9.5 (the environment of Asian adolescents) to 5.5 (the environment of White adolescents) for all adolescents would reduce binge drinking from 6.4% to 5.6% and driving after drinking from 7.9% to 6.0%. Both types of alcohol outlets can contribute to adolescent binge drinking with the same magnitude of effects. As to driving after drinking, the effect of on-sales is larger and statistically significant. However, each of the point estimates for on-sales and off-sales is contained in the confidence interval of the other, so there is no major difference between them. The likelihood that on-sale establishments sell alcohol to underage youth is not much lower than that of off-sale establishments either.19 We therefore conclude that any outlet types within a half mile matter and we do not have the statistical precision to distinguish on-sale and off-sale effects on driving after drinking.
Our findings also confirm sociodemographic disparities in alcohol environments.25−27 Alcohol availability, measured by mean and median number of license, is significantly higher around residences of minority and lower-income families. Some of the descriptive associations are due to the fact that minority and lower-income people tend to live in more densely populated areas. But even after controlling for population density in the census tracts, race and income of individual respondents remain highly significant predictors of number of outlets around their home. Zoning is likely to play an important role not captured by densities and more desirable (and expensive) residences are away from high traffic commercial areas. Interestingly, however, the demographic effect is not limited to residential neighborhoods: Around secondary schools nationwide, the percentage of minority students, especially Asian students positively predicts the number of liquor stores within 400m (1/4 mile) of their schools.33
From an ecological standpoint, higher levels of alcohol outlets and advertising within minority and poorer communities stand in stark contrast to lower rates of alcohol use among minorities. 65.3% of Non-Hispanic White adults report currently drinking compared to 46.6% of African Americans and 51.2% of Hispanics.34 Among youth ages 12−20, prevalence of drinking in the past 30 days is 34.3% for non-Hispanic Whites, 20.2% for non-Hispanic Blacks, and 26.6% for Hispanic.35 Culture seems to be an important factor in keeping lower drinking prevalence among minorities given their higher exposure to alcohol sales. Nevertheless, the mismatch between supply and demand in term of the geographical location of alcohol retails may cause minority and low-income residents to suffer disproportionately from some alcohol-related problems not necessarily from their own drinking. Living nearby alcohol outlets may expose them to risks such as violent crimes,32,36,37 motor vehicle crashes and assault violence,30,38 misdemeanor and felony drunken-driving.39,40
We find no association between family income and any drinking or binge drinking, but higher income youth are more likely to drive after drinking, likely to be due to increased accessibility to motor vehicles. Given the significant findings of binge drinking and driving after drinking, the null finding between outlets and at least one drink in the past 30 days was surprising. One possible explanation is that this level of consumption can occur in a context where adolescents can get alcoholic drinks from their parents’ stock or at the dinner table. Binge drinking - often taking place at parties or within a group of friends - requires larger quantities of alcohol. Youth who binge-drink are also more likely to engage in other problem behaviors, including illegal alcohol purchases. Hispanic youth are more likely to drive after drinking, a finding that is also supported by national statistics41 and a study that shows a positive relationship between alcohol sales to minors and percentage of Hispanic residents.20
This study has important limitations. Observational studies of neighborhood effects are subject to a self-selection bias. Drinkers with certain unobserved or unobservable characteristics can choose to live nearby alcohol outlets (as well as outlets opening in areas of higher demand), thus making it appear that the presence of outlets has a larger effect than it really does without such selection. Controlling for parent’ drinking behavior and focusing on youth who have no influence over their residential location should ameliorate such possible biases at least partially. In a sensitivity analysis, we computed the same alcohol outlet statistics for the sample of households without children, the average alcohol outlets are higher for all sociodemographic groups, suggesting that households with children do sort themselves into neighborhoods with less alcohol sales around their home. There can be many other reasons for the association between sociodemographics and outlet density, varying from zoning regulations to economic factors that affect location decision making by the alcohol retail industry, although this does not affect our results about the association of youth drinking and outlets. Our sample is not big enough to detect interactions between sociodemographic groups and alcohol sales, especially because factors associated with adolescent drinking tend to offset one another. For instance, higher income families are more likely to have children who drive and thus these youth are more likely to be engaged in driving after drinking, even though they are the least exposed to alcohol sales.
Many long-term health behaviors are shaped during youth. Problems that require treatment often do not manifest themselves until much later in life, raising the importance of primary prevention for that age group. The highest prevalence of alcohol dependence in the U.S. population is among people ages 18 to 29,42 who typically started drinking earlier in their teenage years.43 Our study suggests that living in close proximity to alcohol outlets is a risk factor for youth. In California, retail licenses are not typically approved within 100 feet of a residence or within 600 feet of schools, public playgrounds and nonprofit youth facilities, but proximity by itself is not sufficient to deny a license.28 Our study suggests that more attention on the proximity rule is needed and environmental interventions need to curb opportunities for youth to get alcohol from commercial sources, whether being through tightening licensure or enforcing minimum age drinking laws.
Acknowledgments
This research was funded by the Robert Wood Johnson Foundation's Substance Abuse Policy Research Program, grant 63262 and NIAAA grant AA017265. Funding for data access was provided by the National Institute of Environmental Health Sciences, grant P50ES012383.
Contributor Information
Khoa Dang Truong, Pardee RAND Graduate School Email: tdkhoa9@yahoo.com.
Roland Sturm, RAND Corporation Email: sturm@rand.org.
References
- 1.Substance Abuse and Mental Health Services Administration . Results from the 2004 National Survey on Drug Use and Health: National Findings. Rockville, MD: 2005. (Office of Applied Studies, NSDUH Series H-28, DHHS Publication No. SMA 05−4062). [Google Scholar]
- 2.Smith GS, Branas CC, Miller TR. Fatal nontraffic injuries involving alcohol: A metaanalysis. Annals of Emergency Medicine. 1999;33:659–668. [PubMed] [Google Scholar]
- 3.National Highway Traffic Safety Administration (NHTSA) Traffic Safety Facts 2002: Alcohol. NHTSA, National Center for Statistics and Analysis; Washington, DC: 2003. DOT Pub. No. HS–809−606. Available online at: http:///wwwnrd.nhtsa.dot.gov/pdf/nrd30/NCSA/ TSF2002/2002alcfacts.pdf. [Google Scholar]
- 4.Wells S, Graham K, Speechley M, Koval J. Drinking patterns, drinking contexts, and alcohol-related aggression among late adolescent and young adult drinkers. Addiction. 2005;100:933–944. doi: 10.1111/j.1360-0443.2005.001121.x. [DOI] [PubMed] [Google Scholar]
- 5.Cooper ML, Pierce RS, Huselid RF. Substance use and sexual risk taking among black adolescents and white adolescents. Health Psychology. 1994;13:251–262. doi: 10.1037//0278-6133.13.3.251. [DOI] [PubMed] [Google Scholar]
- 6.Cooper ML, Orcutt HK. Drinking and sexual experience on first dates among adolescents. Journal of Abnormal Psychology. 1997;106:191–202. doi: 10.1037//0021-843x.106.2.191. [DOI] [PubMed] [Google Scholar]
- 7.Hingson R, Heeren T, Winter M, et al. Magnitude of alcohol related mortality and morbidity among U.S. college students age 1824: Changes from 1998 to 2001. Annual Review of Public Health. 2005;26:259–279. doi: 10.1146/annurev.publhealth.26.021304.144652. [DOI] [PubMed] [Google Scholar]
- 8.Sher L. Functional magnetic resonance imaging in studies of neurocognitive effects of alcohol use on adolescents and young adults. Int J Adolesc Med Health. 2006 Jan-Mar;18(1):3–7. doi: 10.1515/ijamh.2006.18.1.3. [DOI] [PubMed] [Google Scholar]
- 9.Brown SA, Tapert SF, Granholm E, et al. Neurocognitive functioning of adolescents: Effects of protracted alcohol use. Alcoholism: Clinical and Experimental Research. 2000;24:164–171. [PubMed] [Google Scholar]
- 10.Brown SA, Tapert SF. Adolescence and the trajectory of alcohol use: basic to clinical studies. Ann N Y Acad Sci. 2004 Jun;1021:234–44. doi: 10.1196/annals.1308.028. Review. [DOI] [PubMed] [Google Scholar]
- 11.Crews FT, Braun CJ, Hoplight B, et al. Binge ethanol consumption causes differential brain damage in young adolescent rats compared with adult rats. Alcoholism: Clinical and Experimental Research. 2000;24:1712–1723. [PubMed] [Google Scholar]
- 12.Levy DT, Miller TR, Cox KC. Costs of Underage Drinking. U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention; Washington, DC: 1999. Available online at: http://www.udetc.org/documents/costunderagedrinking.pdf. [Google Scholar]
- 13.Robins LN. Sturdy childhood predictors of adult antisocial behaviour: replications from longitudinal studies. Psychol Med. 1978 Nov;8(4):611–22. doi: 10.1017/s0033291700018821. [DOI] [PubMed] [Google Scholar]
- 14.Gonzalez GM. Early onset of drinking as a predictor of alcohol consumption and alcohol-related problems in college. J Drug Educ. 1989;19(3):225–30. doi: 10.2190/W3RE-MWRR-D8TL-74V0. [DOI] [PubMed] [Google Scholar]
- 15.McGue M, Iacono WG, Legrand LN, Malone S, Elkins I. Origins and consequences of age at first drink. I. Associations with substance-use disorders, disinhibitory behavior and psychopathology, and P3 amplitude. Alcohol Clin Exp Res. 2001 Aug;25(8):1156–65. [PubMed] [Google Scholar]
- 16.U.S. Department of Health and Human Services The Surgeon General's Call to Action To Prevent and Reduce Underage Drinking. U.S. Department of Health and Human Services, Office of the Surgeon General. 2007 [PubMed] [Google Scholar]
- 17.Wagenaar AC, Finnegan JR, Wolfson M, Anstine PS, Williams CL, Perry CL. Where and how adolescents obtain alcoholic beverages. Public Health Reports. 1993;108(4):459–464. [PMC free article] [PubMed] [Google Scholar]
- 18.Forster JL, McGovern PG, Wagenaar AC, Wolfson M, Perry CL, Anstine PS. The ability of young people to purchase alcohol without age identification in northeastern Minnesota, USA. Addiction. 1994 Jun;89(6):699–705. doi: 10.1111/j.1360-0443.1994.tb00956.x. [DOI] [PubMed] [Google Scholar]
- 19.Britt H, Toomey TL, Dunsmuir W, Wagenaar AC. Propensity for and correlates of alcohol sales to Underage Youth. Journal of Alcohol and Drug Education. 2006;50(2):25–42. [Google Scholar]
- 20.Freisthler B, Gruenewald PJ, Treno AJ, Lee J. Evaluating alcohol access and the alcohol environment in neighborhood areas. Alcohol Clin Exp Res. 2003 Mar;27(3):477–84. doi: 10.1097/01.ALC.0000057043.04199.B7. [DOI] [PubMed] [Google Scholar]
- 21.Rhee SH, Hewitt JK, Young SE, et al. Genetic and environmental influences on substance initiation, use, and problem use in adolescents. Archives of General Psychiatry. 2003;60:1256–1264. doi: 10.1001/archpsyc.60.12.1256. [DOI] [PubMed] [Google Scholar]
- 22.Jones-Webb R, Short B, Wagenaar A, Toomey T, Murray D, Wolfson M, Forster J. Environmental predictors of drinking and drinking-related problems in young adults. J Drug Educ. 1997a;27(1):67–82. doi: 10.2190/RJYG-D5C3-H2F0-GJ0L. [DOI] [PubMed] [Google Scholar]
- 23.Jones-Webb R, Toomey TL, Short B, Murray DM, Wagenaar A, Wolfson M. Relationships among alcohol availability, drinking location, alcohol consumption, and drinking problems in adolescents. Subst Use Misuse. 1997b Aug;32(10):1261–85. doi: 10.3109/10826089709039378. [DOI] [PubMed] [Google Scholar]
- 24.Treno AJ, Grube JW, Martin SE. Alcohol availability as a predictor of youth drinking and driving: a hierarchical analysis of survey and archival data. Alcohol Clin Exp Res. 2003 May;27(5):835–40. doi: 10.1097/01.ALC.0000067979.85714.22. [DOI] [PubMed] [Google Scholar]
- 25.LaVeist TA, Wallace JM., Jr. Health risk and inequitable distribution of liquor stores in African American neighborhood. Soc Sci Med. 2000;51(4):613–7. doi: 10.1016/s0277-9536(00)00004-6. [DOI] [PubMed] [Google Scholar]
- 26.Gorman DM, Speer PW. The concentration of liquor outlets in an economically disadvantaged city in the northeastern United States. Subst. Use Misuse. 1997;32:2033–46. doi: 10.3109/10826089709035618. [DOI] [PubMed] [Google Scholar]
- 27.Romley JA, Cohen D, et al. Alcohol and environmental justice: the density of liquor stores and bars in urban neighborhoods in the United States. J Stud Alcohol. 2007;68(1):48–55. doi: 10.15288/jsad.2007.68.48. [DOI] [PubMed] [Google Scholar]
- 28.California Department of Alcoholic Beverage Control [July 09, 2007]; Available at http://www.abc.ca.gov/questions.html.
- 29.Pikora TJ, Bull FC, Jamrozik K, Knuiman M, Giles-Corti B, Donovan RJ. Developing a reliable audit instrument to measure the physical environment for physical activity. Am J Prev Med. 2002 Oct;23(3):187–94. doi: 10.1016/s0749-3797(02)00498-1. [DOI] [PubMed] [Google Scholar]
- 30.Scribner RA, MacKinnon DP, Dwyer JH. Alcohol outlet density and motor vehicle crashes in Los Angeles County cities. Journal of Studies on Alcohol. 1994;55(4):447–453. doi: 10.15288/jsa.1994.55.447. [DOI] [PubMed] [Google Scholar]
- 31.Speer PW, Gorman DM, Labouvie EW, Ontkush MJ. Violent crime and alcohol availability: relationships in an urban community. J Public Health Policy. 1998;19(3):303–18. [PubMed] [Google Scholar]
- 32.Zhu L, Gorman DM, Horel S. Alcohol outlet density and violence: a geospatial analysis. Alcohol Alcohol. 2004 Jul-Aug;39(4):369–75. doi: 10.1093/alcalc/agh062. [DOI] [PubMed] [Google Scholar]
- 33.Sturm R. Disparities in the food environment surrounding US middle and high schools. Public Health. 2008 Jan 17; doi: 10.1016/j.puhe.2007.09.004. [Epub ahead of print] [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Schoenborn CA, Adams PF, et al. “Health Behaviors of Adults: United States, 1999−2001.” National Center for Health Statistics. Vital Health Stat. 2004;10(219) [PubMed] [Google Scholar]
- 35.National Institute on Alcohol Abuse and Alcoholism [November 29, 2007];Surveillance Report #74: Trends in underage drinking in the United States, 1991−2003. Available at http://pubs.niaaa.nih.gov/publications/surveillance74/Underage%2074.pdf.
- 36.Speer PW, Gorman DM, Labouvie EW, Ontkush MJ. Violent crime and alcohol availability: relationships in an urban community. J Public Health Policy. 1998;19(3):303–18. [PubMed] [Google Scholar]
- 37.Gorman DM, Speer PW, Gruenewald PJ, Labouvie EW. Spatial dynamics of alcohol availability, neighborhood structure and violent crime. Journal of Studies on Alcohol. 2001;62(5):628–636. doi: 10.15288/jsa.2001.62.628. [DOI] [PubMed] [Google Scholar]
- 38.Scribner RA, MacKinnon DP, Dwyer JH. The risk of assaultive violence and alcohol availability in Los Angeles. American Journal of Public Health. 1995;85(3):335–340. doi: 10.2105/ajph.85.3.335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Rabow J, Watts RK. Alcohol availability, alcoholic beverage sales and alcohol-related problems. J Stud Alcohol. 1982 Jul;43(7):767–801. doi: 10.15288/jsa.1982.43.767. [DOI] [PubMed] [Google Scholar]
- 40.Watts RK, Rabow J. Alcohol availability and alcohol-related problems in 213 California cities. Alcohol Clin Exp Res. 1983;7(1):47–58. doi: 10.1111/j.1530-0277.1983.tb05410.x. Winter. [DOI] [PubMed] [Google Scholar]
- 41.Youth Risk Behavior Surveillance - United States Morbidity and Mortality Weekly Report. June 28, 2002 / 51(SS04);1−64. [March 29, 2008];Centers for Disease Control and Prevention. 2001 Available at http://www.cdc.gov/mmwr/preview/mmwrhtml/ss5104a1.htm.
- 42.Grant BF, Dawson DA, Stinson FS, Chou SP, Dufour MC, Pickering RP. The 12-month prevalence and trends in DSM-IV alcohol abuse and dependence: United States, 1991−1992 and 2001−2002. Drug Alcohol Depend. 2004 Jun 11;74(3):223–34. doi: 10.1016/j.drugalcdep.2004.02.004. [DOI] [PubMed] [Google Scholar]
- 43.Grant BF, Dawson DA. Age at onset of alcohol use and its association with DSM-IV alcohol abuse and dependence: results from the National Longitudinal Alcohol Epidemiologic Survey. J Subst Abuse. 1997;9:103–10. doi: 10.1016/s0899-3289(97)90009-2. [DOI] [PubMed] [Google Scholar]