Abstract
Objectives. To examine the association between neighborhood socioeconomic status (SES) and alcohol availability before and after deregulation in 2015 of the alcohol market in Ontario, Canada.
Methods. We quantified alcohol access by number of alcohol outlets and hours of retail for all 19 964 neighborhoods in Ontario. We used mixed effects regression models to examine the associations between alcohol access and a validated SES index between 2013 and 2017.
Results. Following deregulation, the number of alcohol outlets in Ontario increased by 15.0%. Low neighborhood SES was positively associated with increased alcohol access: lower-SES neighborhoods had more alcohol outlets within 1000 meters and were closer to the nearest alcohol outlets. Outlets located in low-SES neighborhoods kept longer hours of operation.
Conclusions. We observed a substantial increase in alcohol access in Ontario following deregulation. Access to alcohol was greatest in low-SES neighborhoods and may contribute to established inequities in alcohol harms.
Public Health Implications. Placing limits on number of alcohol outlets and the hours of operation in low-SES neighborhoods offers an opportunity to reduce alcohol-related health inequities.
Alcohol consumption imposes a large burden on human health, contributing to an estimated 6.6% of deaths in men and 2.2% of deaths in women globally.1 In 2015, 22.7 million (77%) Canadians aged 15 years and older drank alcohol in the past year with 4.3 million (20%) and 3.3 million (15%) of these individuals drinking enough to be at risk for chronic and immediate harms, respectively, from their alcohol consumption.2 Alcohol use has a disproportionate impact on low-socioeconomic-status (SES) groups, and research suggests that the same level of alcohol consumption leads to more harms for low-SES individuals than for individuals of higher SES; a finding known as the alcohol harm paradox.3,4
A large body of literature suggests that access to off-premise alcohol outlets, places where alcohol can be purchased only for off-site consumption, is an important contributor to alcohol-associated harms.5–8 Access to outlets may contribute to harmful patterns of alcohol use by facilitating the purchase of alcohol, decreasing the price of alcohol through outlet competition, increasing exposure to marketing and promotion, and making it harder for individuals who have an alcohol use disorder to avoid cues to purchase alcohol. In addition, alcohol outlets appear to concentrate in low-SES neighborhoods, which may contribute to alcohol harms in these communities.9–11
Studies have generally reported a positive association between increased alcohol access and low neighborhood SES. A study in the United States found that census tracts with greater poverty and higher concentrations of African American and Latino populations had more alcohol outlets.9 Another study in Scotland found that low-income neighborhoods had significantly more outlets per capita.10 A study in Quebec, Canada, found that materially deprived (low income and high unemployment) neighborhoods had reduced alcohol access compared with less-deprived neighborhoods, while socially marginalized neighborhoods (higher number of single-parent families and people living alone) had more alcohol access than did less-marginalized neighborhoods.11 However, no studies to our knowledge have investigated the association between neighborhood SES and individual alcohol outlet hours of operation, an important driver of alcohol availability. In addition, most studies have used 1 or 2 dimensions to characterize SES, such as income or ethnicity, and have not accounted for multiple dimensions of SES including social status, income, employment, race/ethnicity, and housing stability.
Over the past 2 decades, several provinces in Canada have deregulated alcohol sales and increased privatization of their alcohol retail markets.12–14 In December 2015, the province of Ontario partially deregulated the alcohol market by allowing grocery stores to apply for a license to sell wine, beer, and cider.15 The purpose of this study was to examine the association between multidimensional measures of neighborhood SES and alcohol access in the context of the partial deregulation of the Ontario alcohol market. This study had 3 objectives: (1) to examine overall alcohol access in Ontario, (2) to examine the impact of partial deregulation on measures of alcohol access, and (3) to examine the relationship between alcohol access and neighborhood SES. We hypothesized that low-SES neighborhoods in Ontario would have greater access to alcohol and that deregulation would result in overall greater alcohol access across the province.
METHODS
Ontario is the largest province in Canada with a population in 2016 of 13.5 million people.16 The sale of alcohol occurs through a mix of public and private retailers, and a government agency, the Alcohol and Gaming Commission of Ontario (AGCO), is responsible for issuing retail licenses. The largest retailer is the Liquor Control Board of Ontario (LCBO), which is operated by the government and sells spirits, wine, and beer. The LCBO licenses “agency stores,” which operate out of existing small businesses in rural communities. The second-largest retailer, The Beer Store, is privately owned and sells beer. Two privately owned retailers, The Wine Rack and The Wine Shop, sell wine. Since December 2015, privately owned grocery stores have been able to apply for a license to sell beer, wine, and cider. Though wineries and breweries may also sell alcohol for consumption off site, we excluded them from analysis because of irregular days of operation and low volume of sales.
Study Design
We used a longitudinal multilevel analysis to examine the association between neighborhood-level SES and alcohol access between 2013 and 2017 (5 time points) across 19 964 Ontario neighborhoods. We chose the Canadian census unit, the Dissemination Area (DA), as being most representative of neighborhoods. The DAs are the smallest geographic unit for which all census data are released and generally contain between 400 and 700 individuals.17
Study Measures
We characterized neighborhood SES by using a validated marginalization index for the province of Ontario, called the Ontario Marginalization Index (OMI).18 The OMI measures 4 dimensions: residential instability, material deprivation, government dependency, and ethnic concentration. Residential instability is a measure of family and housing instability and includes the proportion of the population living alone, the average number of persons per dwelling, the proportion of houses that are not owned, and the proportion of the population who have moved in the past 5 years. Material deprivation measures the ability of individuals in a neighborhood to purchase essential material needs and includes the proportion of the population aged 20 years or older without a high-school degree, lone-parent families, unemployment, and household income. Dependency measures individuals without income from employment and includes the proportion of the population not in the labor force and the number of dependents. Finally, ethnic concentration measures concentrations of recent migrants and visible minority groups (defined by Statistics Canada as “persons, other than aboriginal peoples, who are non-Caucasian in race or non-white in colour”).19 We divided all dimensions into quintiles with a score of 1 representing the least-marginalized quintile and 5 the most-marginalized quintile.
We used 2 measures to control for whether a neighborhood was urban or rural. First, we classified each DA as belonging to a large urban population center (population ≥ 100 000 and population density > 400 people per square kilometer) or outside of a large urban population center by using definitions from Statistics Canada.20 Sixty-one percent (n = 12 335) of DAs were located in a large urban population center. Second, we used the population density of DAs in large urban population centers to further delineate between more urban and suburban areas. We accounted for the total population of each DA along with the percentage of the population that was male and aged 0 to 14, 15 to 64, and 65 or more years, and female aged 0 to 14, 15 to 64, and 65 or more years by using Census data.
A variety of measures have been used in the literature to quantify alcohol access.5,21 We chose to use the Euclidean distance from the geographic center (centroid) of a dissemination area to the nearest alcohol outlet and the number of alcohol outlets located within a Euclidean buffer from the geographic center of a dissemination area as 2 measures of spatial access. We used the total weekly hours of operation for individual alcohol outlets as a measure of temporal access. Because of substantial differences in the meaning of a Euclidean-based distance between a rural and urban region, we decided a priori to limit our spatial analysis to neighborhoods located in large population centers and to pursue interactions in our regression models between population density and each of the 4 OMI dimensions. We retained the interaction in the final model (model 2) if significant.
Alcohol Outlets
Between October and December 2017, we accessed the official Web sites of the LCBO, The Beer Store, The Wine Rack, The Wine Shop, and Government of Ontario to compile a list of all alcohol retail outlets for 2017 (n = 1792). We cross-referenced this list with a list of all active alcohol outlet licenses in Ontario in December 2017, which was provided by the AGCO under a Freedom of Information Request (FOI). The FOI also included the application status (approved, pending, rejected, or cancelled) for all grocery stores that had applied for a license to sell alcohol between December 2015 and July 2018. In October 2017, we extracted the hours of operation for all individual alcohol outlets by accessing the Web sites of individual outlet or retailer Web sites. We used the WayBack machine, an Internet archive resource, to access the historical Web sites of the retailers and compiled a list of alcohol outlets that were open in December for each of the previous 4 years.22 Because no historical Web pages were available for the 159 The Wine Rack stores, we assumed that there had been no changes between 2013 and 2017.
Data Analysis
We conducted all data analysis in Stata version 15.1 (StataCorp LP, College Station, TX) and ArcGIS Pro 2.2 (Environmental Systems Research Institute, Redlands, CA). We used ArcGIS to geocode the addresses of alcohol outlets to a corresponding latitude and longitude. We also used ArcGIS to calculate the Euclidean distance from the geographic center of each dissemination area to the nearest liquor outlet and the number of liquor outlets within buffers of 500 meters (1640 feet) 1000 meters (3281 feet), and 3000 meters (9842 feet) from the geographic center of each DA.
We fit linear regression models for the association between neighborhood marginalization and weekly hours of operation of alcohol outlets in the year 2017. We fit multilevel linear and Poisson models with random effects for each DA to model the association between changes in OMI dimensions and measures of alcohol access over the 5-year period. We modeled each OMI dimension as ordinal quintiles after checking for linearity. For each outcome, we fit 2 models; model A models the overall association of OMI dimension with the outcome, while model B (model C for hours of operation) additionally includes interactions between population density and OMI dimensions determined to be significant. To account for potential clustering of neighborhoods by OMI categories (e.g., where neighborhoods with high levels of material deprivation cluster together creating a larger region of deprivation), we also included a random intercept for membership in a higher-order geographic region (forward sortation area n = 513). We report standard deviations of DA- and forward sortation area–level random effects and computed the intraclass correlation coefficient at the forward sortation area and DA levels for the mixed linear regressions to characterize the degree of clustering. We adjusted all models for population density and the age and gender composition of the DA. We offset the Poisson regressions by the total population of each DA to model the rate of alcohol outlets per capita and interpreted our exponentiated Β coefficients as rate ratios. Finally, we conducted a sensitivity analysis for the Poisson models with 3 different distances: 500, 1000, and 3000 meters.
RESULTS
Figure 1 displays the number of alcohol outlets and alcohol access, by year, from 2013 to 2017. Both the total number of alcohol outlets and spatial measures of alcohol access did not change substantially between 2013 and 2014 (before deregulation). Following deregulation, access to alcohol increased annually. Between 2014 and 2017, the total number of alcohol outlets increased by 15.0%. In 2017, the geographic centers of neighborhoods in large urban population centers were, on average, 55.3 (95% confidence interval = 37.6, 73.2) meters closer to the nearest alcohol outlet than in 2014. Over the same time frame, there was a 20% increase—from 0.96 in 2014 to 1.15 in 2017—in the number of alcohol outlets located within 1000 meters from the geographic center of the average urban neighborhood. Finally, the percentage of privately owned and operated alcohol outlets also increased in the postderegulation time period from 59.0% to 63.2%. Further details on the changes in alcohol stores between 2013 and 2017 can be seen in Table A (available as a supplement to the online version of this article at http://www.ajph.org).
FIGURE 1—
Change in Number of Alcohol Outlets and Number of Grocery Stores Selling Alcohol: Ontario, Canada, 2013–2017
Note. The dashed line separates measurements from the pre- and postderegulation periods. The mean number of alcohol outlets measures the average number of alcohol outlets within 1000 meters of the geographic center of an urban neighborhood in Ontario. Whiskers indicate 95% confidence intervals.
Characteristics of Grocery Stores Selling Alcohol
Newly approved grocery stores selling alcohol (n = 206 in October 2017) represented 88.0% of the observed increase in total outlets. In 2017, grocery stores selling alcohol were open on average for 85.2 hours (94%) of the 91 maximum weekly hours of operation allowed by the government. This weekly total was 19.3 hours greater than the average weekly total of the government-run LCBOs in the same year. According to data from the FOI request, the AGCO received 387 applications from grocery stores for a license to sell alcohol between December 20, 2015, and July 31, 2018. As of July 31, 2018, the AGCO had approved 380 of these applications (98.2%), 2 stores were under review, 5 stores had cancelled their application for a license (with no reason provided), and no applications has been rejected.
Marginalization and Alcohol Availability
Across all years, the geographic center of a neighborhood in a large urban center was on average 1138 meters from the nearest alcohol outlet. Neighborhoods in the most-marginalized quintile of material deprivation, residential instability, and government dependency, respectively, were located 200.5, 301.7, and 173.8 meters closer to a liquor outlet compared with the least-marginalized neighborhoods (Table 1, model A). When we tested for interactions between population density and OMI dimensions, there was a significant interaction with material deprivation where the strength of association between material deprivation and alcohol outlet proximity was stronger in regions of lower population density (Table 1, model B).
TABLE 1—
Mixed Effects Linear Regression Fitting the Distance in Meters From the Geographic Center of Urban Neighborhoods to the Nearest Alcohol Outlet and Socioeconomic Status: Ontario, Canada, 2013–2017
| Model Aa | Model Ba | |
| OMI dimensions,b distance in meters (95% CI) | ||
| Material deprivation | −50.12 (−60.79, −39.44) | −139.25 (−162.85, −115.65) |
| Residential instability | −75.43 (−86.15, −64.72) | −68.09 (−91.94, −44.24) |
| Ethnic concentration | 20.28 (6.89, 33.67) | 18.44 (3.81, 33.06) |
| Government dependency | −43.44 (−55.34, −31.54) | −44.74 (−57.02, −32.36) |
| Population density, quintilesc | −67.35 (−77.12, −57.59) | −144.98 (−169.07, −120.88) |
| Total population, hundreds | 5.21 (3.61, 6.82) | 4.95 (3.35, 6.55) |
| Year (Ref = 2013), distance in meters (95% CI) | ||
| 2014 | −2.74 (−6.45, 0.96) | −2.74 (−6.45, 0.96) |
| 2015 | −17.81 (−21.51, −14.10) | −17.81 (−21.51, −14.10) |
| 2016 | −26.44 (−30.15, −22.74) | −26.44 (−30.15, −22.74) |
| 2017 | −58.15 (−61.86, −54.45) | −58.15 (−61.86, −54.45) |
| Intercept | 2364.08 (2162.82, 2565.35) | 2565.30 (2345.65, 2784.96) |
| Interaction: material deprivation × population density | 26.57 (20.33, 32.82) | |
| Mixed effects, SD | ||
| Residuals | 146.72 | 146.73 |
| Forward sortation area | 532.53 | 525.77 |
| Neighborhood (DA) | 528.04 | 526.85 |
| Intraclass correlation coefficients (95% CI) | ||
| Forward sortation area | 0.49 (0.44, 0.53) | 0.48 (0.43, 0.53) |
| Neighborhood (DA) | 0.96 (0.96, 0.97) | 0.96 (0.96, 0.97) |
Note. CI = confidence interval; DA = Dissemination Area; OMI = Ontario Marginalization Index.
Both model A and model B are adjusted for the percentage of the male population who are aged 15 to 65 years and 65 years or older and the percentage of the female population who are aged 15 to 65 years and 65 years or older. Mixed effects are for each individual neighborhood (DA) and membership in a larger geographic area (forward sortation area).
The coefficients of the 4 OMI dimensions represent the effect of increasing the marginalization of a neighborhood by 1 quintile on the distance to the nearest alcohol outlet.
The coefficient for population density represents the effect of a quintile increase in population density on the distance to the nearest alcohol outlet.
Each quintile increase in material deprivation and residential instability was associated with a statistically significant 15% increase in the number of outlets 1000 meters from the geographic center of a neighborhood and a quintile increase in government dependency was associated with a 20% increase in the number of outlets (Table 2, model A). Conversely, each quintile increase in ethnic concentration was associated with 18% fewer alcohol outlets. When we included significant interactions between population density and OMI dimensions, there was a significant interaction between population density and material deprivation and population density and residential instability. While higher material deprivation and residential instability were associated with more alcohol outlets located within 1000 meters, this effect was more pronounced in regions of low population density compared with higher population density (Table 2, model B). In addition, the sensitivity analysis showed a consistent association for the number of outlets within different distances (500, 1000, and 3000 meters) and SES (Table B, available as a supplement to the online version of this article at http://www.ajph.org).
TABLE 2—
Mixed Effects Poisson Model Fitting the Number of Alcohol Outlets Within 1000 Meters From the Geographic Center of Urban Neighborhoods and Socioeconomic Status: Ontario, Canada, 2013–2017
| Model A,a RR (95% CI) or SD | Model B,a RR (95% CI) or SD | |
| OMI dimensionsb | ||
| Material deprivation | 1.15 (1.10, 1.19) | 1.59 (1.45, 1.73) |
| Residential instability | 1.15 (1.10, 1.20) | 1.35 (1.22, 1.49) |
| Ethnic concentration | 0.82 (0.79, 0.86) | 0.84 (0.81, 0.88) |
| Government dependency | 1.20 (1.15, 1.25) | 1.21 (1.16, 1.26) |
| Population density, quintilesc | 1.27 (1.22, 1.32) | 1.98 (1.80, 2.18) |
| Year (Ref = 2013) | ||
| 2014 | 0.99 (0.97, 1.02) | 0.99 (0.97, 1.02) |
| 2015 | 1.07 (1.04, 1.09) | 1.06 (1.04, 1.09) |
| 2016 | 1.10 (1.07, 1.13) | 1.10 (1.07, 1.13) |
| 2017 | 1.19 (1.16, 1.22) | 1.19 (1.16, 1.22) |
| Interactions | ||
| Material deprivation × population density | . . . | 0.91 (0.89, 0.93) |
| Residential instability × population density | . . . | 0.96 (0.94, 0.98) |
| Mixed effects | ||
| Forward sortation area | 1.19 | 1.19 |
| Neighborhood (DA) | 1.71 | 1.71 |
Note. CI = confidence interval; DA = Dissemination Area; OMI = Ontario Marginalization Index; RR = rate ratio.
Both model A and model B are offset by the total population of a dissemination area and adjusted for the percentage of the male population who are aged 15 to 65 years and 65 years or older and the percentage of the female population who are aged 15 to 65 years and 65 years or older. Mixed effects are for each individual neighborhood (DA) and membership in a larger geographic area (forward sortation area).
The coefficients of the 4 OMI dimensions represent the effect of increasing the marginalization of a neighborhood by 1 quintile on the number of alcohol outlets within 1000 meters.
The coefficient for population density represents the effect of a quintile increase in population density on the number of alcohol outlets within 1000 meters.
Alcohol outlets located in neighborhoods in the most-marginalized quintile of residential instability and ethnic concentration were open 2.12 and 3.28 more hours per week, respectively, compared with outlets located in the least-marginalized neighborhoods (Table 3 model B). Outlets located in areas of high population density kept longer hours of operation. When we included significant interactions between population density and OMI dimensions, there was a significant interaction between population density and material deprivation. In regions with lower population density, outlets located in materially deprived neighborhoods kept relatively longer hours than outlets in less materially deprived neighborhoods. However, this trend was less pronounced in neighborhoods in higher-population-density areas (Table 3, model C). Each additional outlet in a neighborhood in a large urban center was associated with an increase of 1.4 weekly hours. The hours of operation differed by outlet type; wine stores and grocery stores were open significantly longer than the LCBO and The Beer Stores.
TABLE 3—
Linear Regression Model Fitting Individual Outlet Weekly Hours of Operation With Neighborhood Socioeconomic Status and Outlet Type: Ontario, Canada, 2013–2017
| Model A,a Change in Hours (95% CI) | Model B,b Change in Hours (95% CI) | Model C,b Change in Hours (95% CI) | |
| OMI dimensionsc | |||
| Material deprivation | . . . | 0.01 (−0.31, 0.34) | 1.39 (0.60, 2.18) |
| Residential instability | . . . | 0.13 (−0.25, 0.52) | 0.11 (−0.27, 0.49) |
| Ethnic concentration | . . . | 0.82 (0.39, 1.24) | 0.87 (0.44, 1.29) |
| Government dependency | . . . | 0.53 (0.08, 0.98) | 0.59 (0.14, 1.04) |
| Population density, quintilsd | . . . | 0.36 (0.01, 0.71) | 1.96 (1.05, 2.87) |
| No. of outlets | . . . | 1.42 (0.88, 1.96) | 1.35 (0.81, 1.89) |
| Total population, hundreds | . . . | −0.01 (−0.05, 0.03) | −0.01 (−0.05, 0.03) |
| Type of outlet (Ref = LCBO) | |||
| The Beer Store | 1.36 (−0.13, 2.84) | −0.63 (−1.66, 0.40) | −0.67 (−1.69, 0.35) |
| The Wine Shop | 9.62 (7.00, 12.23) | 2.80 (1.32, 4.27) | 2.54 (1.07, 4.00) |
| The Wine Rack | 11.07 (8.91, 13.23) | 5.75 (4.49, 7.02) | 5.67 (4.42, 6.93) |
| Agency storee | 0.09 (−1.86, 2.03) | NA | NA |
| Grocery store | 19.26 (17.32, 21.2) | 13.17 (11.95, 14.38) | 12.97 (11.77, 14.18) |
| Intercept | 65.89 (64.67, 67.11) | 62.27 (54.99, 69.55) | 57.57 (49.93, 65.20) |
| Interaction: population density × deprivation | . . . | . . . | −0.45 (−0.69, −0.21) |
| Adjusted R2 | 0.275 | 0.466 | 0.475 |
Notes. CI = confidence interval; LCBO = Liquor Control Board of Ontario; NA = not applicable.
Model A shows the unadjusted average hours by outlet type for all of Ontario.
Models B and C are for large urban population centers only and are adjusted for the percentage of the male population who are aged 15 to 65 years and 65 years or older and the percentage of the female population who are aged 15 to 65 years and 65 years or older.
The coefficients of 4 OMI dimensions represent the effect of increasing the marginalization of a neighborhood by 1 quintile on the total weekly hours of operation for an outlet in that neighborhood.
The coefficient for population density represents the effect of increasing the population density of a neighborhood by 1 quintile on the total weekly hours of operation for an outlet in that neighborhood.
LCBO agency stores were only located outside large urban population centers.
DISCUSSION
Deregulation of the alcohol market in 2015 resulted in increased alcohol availability in Ontario, including increased number of outlets, proximity to outlets, and weekly hours of operation. We observed significant associations between neighborhood SES and access to alcohol outlets. Neighborhoods that were more materially deprived, government dependent, and residentially instable were located closer to alcohol outlets and had higher numbers of alcohol outlets within 1000 meters. Conversely, neighborhoods with more visible minorities and migrants were located farther from alcohol outlets and had fewer outlets within 1000 meters. Outlets that were located in neighborhoods with more visible minorities and migrants, with more individuals who were materially deprived or with more individuals on government assistance, kept longer weekly hours of operation.
The observed increase in physical access to alcohol following deregulation was an expected result of increasing the number of alcohol outlets. However, deregulation also resulted in increased of hours of alcohol sales across Ontario as grocery stores kept longer hours of operation than did all other types of alcohol outlets. We speculate that grocery stores keep longer hours (94% of the legal maximum) because they sell products other than alcohol. As a consequence, the marginal cost of keeping the store open is lower. Deregulation also resulted in an increased proportion of the alcohol market in Ontario being privately operated. Several studies have found that increases in private alcohol outlets are associated with higher per-capita alcohol sales, and private outlets may be more likely to sell alcohol to minors and less likely to comply with age verifications.23,24
We found a positive association between increased alcohol access and material deprivation, residential instability, and reliance on government assistance. There are several possible explanations for these findings. First, commercial zoning bylaws may result in concentration of alcohol outlets in low-SES neighborhoods. Second, low-SES neighborhoods may have higher demand for alcohol and market forces draw outlets to these areas. Third, lower prices to purchase or rent storefronts in low-SES neighborhoods may attract alcohol retailers. A recent study in Ontario found that tobacco retailers across the province concentrate in low-SES neighborhoods and near schools.25 Similar factors may be driving the placement of both tobacco and alcohol retailers. Conversely, neighborhoods with a higher number of immigrants and visible minorities had less alcohol availability. This finding is consistent with research showing that individuals born outside of Canada consume less alcohol than do Canadian-born individuals, suggesting that outlets may be responding to lower market demand.26
Our FOI request revealed that the AGCO had a permissive standard to grocery stores’ alcohol license applications. While the AGCO lists 6 criteria when considering an application for a grocery store to sell alcohol (Box A, available as a supplement to the online version of this article at http://www.ajph.org), these criteria do not consider proximity to other alcohol outlets (excluding LCBO Agency Stores) or whether communities are at risk for alcohol-related harms.27 Our results suggest that overall AGCO licensing criteria have allowed high concentrations of alcohol access in vulnerable neighborhoods.
This is the first study to our knowledge to examine the association between individual alcohol outlet hours and neighborhood SES. We observed that outlets in low-SES neighborhoods and some types of privately run outlets kept longer hours of operation than did high-SES neighborhood or government outlets. We also found that outlets in close proximity to one another kept longer hours of operation, which we speculate may be the result of competition between outlets. While further research is needed to understand the driving factors behind these observations, previous research is suggestive that increased daily hours of operation could result in more alcohol consumption and harms.28,29 Collectively, the higher concentration of alcohol outlets and longer hours of operation in lower-SES neighborhoods are factors that likely contribute to disparities in alcohol harms. Individuals living in low-SES neighborhoods who are exposed to high availability of alcohol retail may suffer harm directly through increased alcohol consumption and indirectly arising from the behaviors of other individuals consuming alcohol, such as road traffic crashes or assaults.6
Strengths and Limitations
There are several strengths to this study. We used a validated marginalization index to investigate the association between several SES dimensions. We considered both physical access to alcohol outlets and access to hours of sale. The study design measured alcohol access longitudinally across a large number of stable geographic regions. Finally, we used multilevel modeling to control for potential impacts of spatial autocorrelation on our findings, which could decrease confounding attributable to clustering of marginalized neighborhoods.
Nonetheless, several important limitations must be noted. First, increased access to alcohol is an important public health concern if it contributes to increased harmful patterns of alcohol consumption or alcohol-related harms. It is possible that the observed increase in alcohol access will result in individuals purchasing alcohol from different locations without increasing their overall consumption or alcohol-related harms. While we did not directly investigate the relationship between increased alcohol availability and harms in this study, a large body of literature suggests that increasing alcohol access will result in these negative outcomes.5–8 Future research should investigate the impact of alcohol deregulation in Ontario on alcohol consumption and associated harms.
Second, we examined only off-premise sales locations and did not account for alcohol access through on-premise locations, such as bars and restaurants. It is possible that neighborhoods with lower access to off-premise alcohol retailers had other opportunities to access alcohol through bars and restaurants.
Third, because there is no information about the distribution of individuals within each neighborhood, we assumed that the geographic center of a neighborhood represented the location of the “average individual.” This assumption is likely a reasonable approximation for urban neighborhoods, which are small and uniform in shape, but may misrepresent individual distribution in rural or suburban areas where neighborhoods are irregularly shaped and may include uninhabited regions. Although we chose to limit our geospatial analyses to large, urban population centers, these areas still include some suburbs.
Fourth, both spatial measures of alcohol availability used Euclidean distances. In small geographic regions, these estimates are accurate but may lose accuracy in larger areas with less uniform road networks. More sophisticated measures of alcohol availability that capture travel opportunity, cost, and time, in addition to distance, would allow generalization to all regions of the province.
Finally, we used the OMI values from the 2011 census year for all times in our study. Although marginalization is likely relatively stable over a 5-year time period, it is possible that these measures have changed over time. Future work could examine both longitudinal changes in alcohol access and marginalization.
CONCLUSIONS
In the absence of regulatory oversight, alcohol access appears to concentrate in low-SES neighborhoods. Deregulation of the Ontario alcohol market has resulted in an increase in the number of locations and hours of operation to purchase alcohol. Further research is needed to understand the impact of increased alcohol access on alcohol consumption and alcohol-related harms in Ontario. Regulations to reduce alcohol-related inequities could consider addressing both current availability of alcohol and neighborhood socioeconomic factors when the location of future alcohol retail outlets is being determined.
ACKNOWLEDGMENTS
We thank Susan Mowers and Rene Duplain, data librarians at the University of Ottawa, for their assistance accessing data from the Canadian Census. John Pearson helped retrieve the hours of operation for the grocery stores in Ontario.
CONFLICTS OF INTEREST
All authors have no competing interests to disclose.
HUMAN PARTICIPANT PROTECTION
Given that this study used de-identified, aggregate-level data, it was deemed exempt from institutional review board approval (Harvard institutional review board, IRB17-1981).
Footnotes
REFERENCES
- 1.GBD 2016 Alcohol Collaborators. Alcohol use and burden for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet. 2018;392(10152):1015–1035. doi: 10.1016/S0140-6736(18)31310-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Government of Canada. Canadian Tobacco Alcohol and Drugs (CTADS): 2015 summary. 2017. Available at: https://www.canada.ca/en/health-canada/services/canadian-tobacco-alcohol-drugs-survey/2015-summary.html. Accessed February 12, 2018.
- 3.Jones L, Bates G, McCoy E, Bellis MA. Relationship between alcohol-attributable disease and socioeconomic status, and the role of alcohol consumption in this relationship: a systematic review and meta-analysis. BMC Public Health. 2015;15(1):400. doi: 10.1186/s12889-015-1720-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Katikireddi SV, Whitley E, Lewsey J, Gray L, Leyland AH. Socioeconomic status as an effect modifier of alcohol consumption and harm: analysis of linked cohort data. Lancet Public Health. 2017;2(6):e267–e276. doi: 10.1016/S2468-2667(17)30078-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Holmes J, Guo Y, Maheswaran R, Nicholls J, Meier PS, Brennan A. The impact of spatial and temporal availability of alcohol on its consumption and related harms: a critical review in the context of UK licensing policies. Drug Alcohol Rev. 2014;33(5):515–525. doi: 10.1111/dar.12191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Popova S, Giesbrecht N, Bekmuradov D, Patra J. Hours and days of sale and density of alcohol outlets: impacts on alcohol consumption and damage: a systematic review. Alcohol Alcohol. 2009;44(5):500–516. doi: 10.1093/alcalc/agp054. [DOI] [PubMed] [Google Scholar]
- 7.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(6):556–569. doi: 10.1016/j.amepre.2009.09.028. [DOI] [PubMed] [Google Scholar]
- 8.Stockwell T, Zhao J, Martin G et al. Minimum alcohol prices and outlet densities in British Columbia, Canada: estimated impacts on alcohol-attributable hospital admissions. Am J Public Health. 2013;103(11):2014–2020. doi: 10.2105/AJPH.2013.301289. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Berke EM, Tanski SE, Demidenko E, Alford-Teaster J, Shi X, Sargent JD. Alcohol retail density and demographic predictors of health disparities: a geographic analysis. Am J Public Health. 2010;100(10):1967–1971. doi: 10.2105/AJPH.2009.170464. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Shortt NK, Tisch C, Pearce J et al. A cross-sectional analysis of the relationship between tobacco and alcohol outlet density and neighbourhood deprivation. BMC Public Health. 2015;15(1):1014. doi: 10.1186/s12889-015-2321-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ngamini Ngui A, Apparicio P, Philibert M, Fleury M-J. Neighborhood characteristics associated with the availability of alcohol outlets in Quebec. J Addict. 2015;2015:876582. doi: 10.1155/2015/876582. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Trolldal B. An investigation of the effect of privatization of retail sales of alcohol on consumption and traffic accidents in Alberta, Canada. Addiction. 2005;100(5):662–671. doi: 10.1111/j.1360-0443.2005.01049.x. [DOI] [PubMed] [Google Scholar]
- 13.Stockwell T, Zhao J, Marzell M et al. Relationships between minimum alcohol pricing and crime during the partial privatization of a Canadian government alcohol monopoly. J Stud Alcohol Drugs. 2015;76(4):628–634. doi: 10.15288/jsad.2015.76.628. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Slaunwhite AK, McEachern J, Ronis ST, Peters PA. Alcohol distribution reforms and school proximity to liquor sales outlets in New Brunswick. Can J Public Health. 2018;108(5–6):488. doi: 10.17269/CJPH.108.6132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Government of Ontario. Initial report: the Premier’s Advisory Council on Government Assets. 2015. Available at: https://www.ontario.ca/page/initial-report-premiers-advisory-council-government-assets#section-3. Accessed May 14, 2018.
- 16.Statistics Canada. 2017. Ontario [Province] and Canada [Country] (table). Census Profile. 2016 Census. Statistics Canada Catalogue no. 98-316-X2016001. Ottawa. Released November 29, 2017. Available at: http://www12.statcan.gc.ca/census-recensement/2016/dp-pd/prof/details/page.cfm?Lang=E&Geo1=PR&Code1=35&Geo2=PR&Code2=01&Data=Count&SearchText=ontario&SearchType=Begins&SearchPR=01&B1=All&TABID=1. Accessed February 9, 2018.
- 17.Statistics Canada. Dissemination area (DA). Census Dictionary. 2015. Available at: https://www12.statcan.gc.ca/census-recensement/2011/ref/dict/geo021-eng.cfm. Accessed September 30, 2018.
- 18.Matheson FI, Dunn JR, Smith KL, Moineddin R, Glazier RH. Development of the Canadian Marginalization Index: a new tool for the study of inequality. Can J Public Health. 2012;103(suppl 2):S12–S16. doi: 10.1007/BF03403823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Statistics Canada. Visible minority of person. 2015. Available at: http://www23.statcan.gc.ca/imdb/p3Var.pl?Function=DEC&Id=45152. Accessed March 27, 2019.
- 20.Statistics Canada. Population centre and rural area classification 2016. 2016. Available at: https://www.statcan.gc.ca/eng/subjects/standard/pcrac/2016/introduction. Accessed September 10, 2018.
- 21.Centers for Disease Control and Prevention. Guide for measuring alcohol outlet density. US Department of Health and Human Services. 2017. Available at: https://www.cdc.gov/alcohol/pdfs/CDC-Guide-for-Measuring-Alcohol-Outlet-Density.pdf. Accessed December 2, 2018.
- 22.Arora SK, Li Y, Youtie J, Shapira P. Using the wayback machine to mine websites in the social sciences: a methodological resource. J Assoc Inf Sci Technol. 2016;67(8):1904–1915. [Google Scholar]
- 23.Giesbrecht N, Bosma LM. Preventing Alcohol-Related Problems: Evidence and Community-Based Initiatives. Washington, DC: American Public Health Association; 2017. Preventing alcohol-related problems. [Google Scholar]
- 24.Hahn RA, Middleton JC, Elder R et al. Effects of alcohol retail privatization on excessive alcohol consumption and related harms. Am J Prev Med. 2012;42(4):418–427. doi: 10.1016/j.amepre.2012.01.002. [DOI] [PubMed] [Google Scholar]
- 25.Chaiton MO, Mecredy GC, Cohen JE, Tilson ML. Tobacco retail outlets and vulnerable populations in Ontario, Canada. Int J Environ Res Public Health. 2013;10(12):7299–7309. doi: 10.3390/ijerph10127299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Agic B, Mann RE, Tuck A et al. Alcohol use among immigrants in Ontario, Canada. Drug Alcohol Rev. 2016;35(2):196–205. doi: 10.1111/dar.12250. [DOI] [PubMed] [Google Scholar]
- 27.Alcohol and Gaming Commission of Ontario. Authorization to sell beer and wine or beer and cider in a grocery store. 2018. Available at: http://agco.ca/sites/default/files/1577e_guide_beer_grocerystores_0.pdf. Accessed August 17, 2018.
- 28.Sherk A, Stockwell T, Chikritzhs T et al. Alcohol consumption and the physical availability of take-away alcohol: systematic reviews and meta-analyses of the days and hours of sale and outlet density. J Stud Alcohol Drugs. 2018;79(1):58–67. [PubMed] [Google Scholar]
- 29.Canadian Public Health Association. Too high a cost. A public health approach to alcohol policy in Canada. 2011. Available at: https://www.cpha.ca/sites/default/files/assets/positions/position-paper-alcohol_e.pdf. Accessed February 7, 2018.