3.1. Results
See
Figure 1 for the distribution of outlets across the province. Convenience stores were the most common type of outlet (56%), followed by gas stations (20%), grocery stores (9%), and bars (6%). Many retail outlets were located within walking distance of each other. In urban areas, 86% and 73% were located with 500 m and 250 m, respectively, of another retailer (data not shown). Even in rural areas, 63% and 52% of tobacco retail outlets were located with 500 m and 250 m of another retailer, respectively.
This corresponds to approximately one location per 1,000 people aged 15+, five locations per 1,000 children <15, or five locations per 1,000 smokers. Most outlets were located in the major population centers (
Table 1) with the proportion of retailers reflecting the proportion of the population; however, outlets were slightly over represented in small towns (population of less than 10,000) (16% of retailers compared to 12% of the Ontario population).
Density of outlets per 1,000 people over 15 years old varied by the 36 public health units ranging from a low of 0.09 to a high of 1.65 with the average density of 1.05 per 1,000 people over 15 (SD (0.35 per 1,000 people over 15). Increased density was associated with higher prevalence of smoking within a public health unit (r = 0.36, r2 = 0.13).
Figure 1.
Map of tobacco retailers in Ontario, with enlarged view of southern Ontario.
Figure 1.
Map of tobacco retailers in Ontario, with enlarged view of southern Ontario.
Table 1.
Distribution of Tobacco Retail Outlets compared to the Ontario population.
Table 1.
Distribution of Tobacco Retail Outlets compared to the Ontario population.
| Ontario Tobacco Retailers | Ontario Population |
|---|
| % | % |
| Area classification type * | | |
| Census metropolitan area | 74 | 79 |
| Census agglomeration | 10 | 9 |
| Non-metropolitation/agglomeration | 16 | 12 |
| Rurality | | |
| Urban | 84 | 85 |
| Rural | 16 | 15 |
| Community size | | |
| 1,500,000 + | 42 | 42 |
| 500,000–1,499,999 | 10 | 12 |
| 100,000–499,999 | 23 | 25 |
| 10,000–99,999 | 9 | 9 |
| <10,000 | 16 | 12 |
In urban areas, tobacco retail outlets were more likely to be located in neighbourhoods that had higher deprivation, but there were fewer outlets in neighbourhoods with a high percentage of immigrants, independently of income level and neighbourhood deprivation (
Table 2). There was no independent effect of the neighbourhood being predominately blue collar. In rural areas, there was similarly no correlation between retailer density and blue collar neighbourhoods, but tobacco outlets were more often located in more deprived neighbourhoods and, unlike in urban areas, neighbourhoods with higher immigrant proportion. Tobacco outlets were more likely to be found in neighbourhoods that had a higher proportion of residents receiving government assistance, being single parent families, having houses needing major repairs, and lacking a university or college degree. Unemployment rates were not associated independently with retail tobacco outlet location.
Table 2.
Odds of having a tobacco retail outlet by neigbourhood (dissemination area) characteristics. (n = 18,922 dissemination areas).
Table 2.
Odds of having a tobacco retail outlet by neigbourhood (dissemination area) characteristics. (n = 18,922 dissemination areas).
| | Urban | Rural |
|---|
| Odds Ratio | p-value | Odds Ratio | p-value |
|---|
| Neighbourhood deprivation * |
| Quartile 1—least deprivation | Referent | | Referent | |
| Quartile 2 | 1.58 | <0.0001 | 3.15 | <0.0001 |
| Quartile 3 | 2.04 | <0.0001 | 3.60 | <0.0001 |
| Quartile 4—most deprivation | 3.14 | <0.0001 | 2.70 | <0.0001 |
| Immigrant population |
| Quartile 1—fewest immigrants | Referent | | Referent | |
| Quartile 2 | 0.99 | 0.77 | 1.61 | 0.0002 |
| Quartile 3 | 0.88 | 0.01 | 1.52 | 0.2556 |
| Quartile 4—most immigrants | 0.89 | 0.01 | | |
| Blue-collar workers † |
| Quartile 1—fewest blue-collar | Referent | | Referent | |
| Quartile 2 | 1.23 | <0.0001 | 2.387 | <0.0001 |
| Quartile 3 | 1.41 | <0.0001 | 2.304 | <0.0001 |
| Quartile 4—most blue-collar | 1.42 | <0.0001 | 2.408 | <0.0001 |
Most retail outlets were located within walking distance of a school (
Table 3). In urban areas, 68% of stores were located within 500 m of a school. In rural areas, 38% were located within 500 m of a school. Schools in lower socio-economic areas were more likely to have a retailer within walking distance (
p < 0.001) (
Table 4). For retailers with residents in the top quintile of income, 25% had schools within 250 m, and 55% had schools within 500 m. For retailers located in areas in the lowest income quintile, 33% had schools within 250 m and 73% had schools within 500 m.
Table 3.
Number of retailers within walking distance of a school.
Table 3.
Number of retailers within walking distance of a school.
| | Rural Retailers
N (%) | Urban Retailers
N (%) | Total
N (%) |
|---|
| #of schools within 500 m of retailer |
| 0 | 1,335 (61.7%) | 2,835 (31.7%) | 4,170 (37.5%) |
| 1+ | 828 (38.3%) | 6,115 (68.3%) | 6,943 (62.5%) |
| #of schools within 250 m of retailer |
| 0 | 1,791 (82.8%) | 6,244 (70%) | 8,035 (72.3%) |
| 1+ | 372 (17.2%) | 2,706 (30%) | 3,078 (27.7%) |
| Total N | 2,163 | 8,950 | 11,113 |
Table 4.
Percent of retailers with one or more schools within walking distance (250 or 500 m), by neighbourhood income quartile (n = 11,013 retailers *).
Table 4.
Percent of retailers with one or more schools within walking distance (250 or 500 m), by neighbourhood income quartile (n = 11,013 retailers *).
| Neighbourhood Income Quintiles | # of Retailers within Meters of a School N (%) |
|---|
| Retailers within 250 m | Retailers within 500 m |
|---|
| 1 (lowest) | 1,025 (33.3%) | 2,236 (72.7%) |
| 2 | 754 (29.2%) | 1,707 (66.2%) |
| 3 | 491 (23.6%) | 1,159 (55.7%) |
| 4 | 392 (22.3%) | 950 (54.0%) |
| 5 (highest) | 385 (25.4%) | 831 (54.9%) |
| p-trend | <0.0001 | <0.0001 |
3.2. Discussion
Availability of tobacco remains widespread across Ontario, as it does throughout most of the world. The number of retail outlets, approximately one for every thousand people older than 15, makes tobacco extremely available. Despite the tremendous burden of disease, disability, and death caused by tobacco, the widespread availability is similar to that of a normal, benign product.
A previous Australian study using retail license data found an outlet density of two per 1,000 in New South Wales [
23] while a study that used retail classification to estimate tobacco outlets found an average of 0.5 outlets per 1,000 people [
24]. The density of tobacco in Ontario is slightly understated as it does not include multiple outlets per location. The number of outlets appears to have fallen in Ontario from a previous estimate of 14,500 outlets based on 2006 data [
15]. Furthermore, while availability of tobacco is ubiquitous, availability is even more prevalent in vulnerable neighborhoods. Retail outlets were more likely to be present in areas that were low income. This level of increased exposure is likely to contribute to the perception that tobacco products are socially acceptable, which increases the likelihood of tobacco uptake and increases the difficulty in both making and succeeding in a quit attempt [
25,
26]. These findings are consistent with studies across North America that have found a relationship between socio-economic status and availability of tobacco products [
27]. The relationship observed in this study is constant across rural and urban areas, suggesting that density of population, per se, is not a factor here.
More surprising, and unlike previous studies, the relationship between immigration level in a neighborhood and tobacco retailer density differed by rurality. Consistent with American literature [
14,
15,
16,
17], rural areas with a high proportion of immigrants were more likely to have greater availability of tobacco; however, the reverse was true in urban areas. Immigrants in Canada are less likely to smoke than the general population, which may suggest that neighborhoods with more immigrants may provide an additional protective effect preventing second generation immigrants from starting to smoke [
28].
Similarly, there was a modest relationship between the density of outlets and the number of smokers, with the density increasing with the prevalence of smoking within public health units. This is likely the result of two main processes: tobacco outlets follow cigarette demand, and the number of smokers is kept high by the increased supply of cigarettes. While it is impossible to distinguish these effects with this cross-sectional analysis, Novak
et al., found that the relationship between outlet density and population increased in magnitude with additional control for potential confounders [
29]. Our analysis also uses a large unit of population size to assess smoking status, similar to or larger than the county level analysis of Peterson
et al. [
19]. Given the limitations of using public health unit as a geographic unit, any association seen in this study would likely be smaller in magnitude than if a smaller, more homogeneous geographical unit had been available to assess the relationship of smoking prevalence and outlet density.
Curiously, there was no relationship between the blue collar status of the neighborhood and the number of outlets, in either rural or urban areas. Both low income adequacy and blue-collar status are known to be consistently associated with increased levels of smoking in Ontario [
30]; however, the differentiation with respect to the presence of tobacco outlets suggests that a mechanism exists other than simple correlation between the number of smokers and the number of outlets. Research from other areas shows that higher density of fast food and alcohol outlets in vulnerable neighborhoods results in corresponding higher rates of alcohol-related problems and overweight/obesity [
31,
32].
Many children have availability to tobacco within a short distance of their schools. A cross-sectional analysis by Henriksen
et al., found that schools located in neighborhoods with a higher proportion of Hispanics and residents of lower socio-economic status were more likely to have higher tobacco retailer density [
18]. Associations between availability of tobacco and smoking prevalence have also been found in Ontario [
33,
34]. Nearly three quarters of retailers in low income areas have a school within a 5 min walk (
i.e., 500 m). This proximity to tobacco retailers serves to increase tobacco exposure and the opportunity to purchase tobacco products [
35,
36,
37] which may aid in the transition from experimenting with smoking to becoming regular smokers [
34].
