CC BY-NC-ND
0Contents lists available at ScienceDirect
Addictive Behaviors
ADDICTIVE BEHAVIORS
Youth retail tobacco access in Canada: Regional variation, perceptions, and predictors from YSS 2010/2011
Leia M. Minaker a* Shilpa Sonib, Nghia Nguyen a, Steve Manske a
a Propel Centre for Population Health Impact, University of Waterloo, Waterloo, Ontario, Canada b School ofPublic Health and Health Systems, University ofWaterloo, Waterloo, Ontario, Canada
HIGHLIGHTS
• A quarter of Canadian grades 9-12 smokers usually buy cigarettes at a store
• Provincial youth retail tobacco access ranges from 16% in BC to 26% in QC
• 80% of never smokers thought it would be easy to get cigarettes to smoke
CrossMark
ARTICLE INFO
Article history: Received 9 December 2014 Received in revised form 17 June 2015 Accepted 30 June 2015 Available online 4 July 2015
Keywords:
Tobacco access
Smoking
Adolescents
Public policy
Surveillance
Canada
Youth smoking survey
ABSTRACT
Purpose: Retail tobacco access is an important determinant of youth smoking prevalence. This study examines perceptions of ease in obtaining cigarettes and how prevalence of self-reported retail tobacco access among youth smokers varies by province in Canada. Additionally, relevant retail experiences, such as being asked for identification by a store clerk, are described.
Methods: Data from grades 9-12 students who participated in the 2010/2011 Youth Smoking Survey, a nationally generalizable sample of Canadian students (n = 31396) were used to examine retail tobacco access and related experiences. Logistic regression models were used to examine differences in retail tobacco access and retail tobacco experiences by sociodemographic and regional characteristics.
Results: 79% of students who never smoked thought it would be easy to get cigarettes. About one-quarter of smokers reported usually buying cigarettes from stores, and the percent of student smokers usually buying cigarettes in stores ranged from 16% in British Columbia to 36% in Quebec. Compared to grade 9 students, grade 12 students had higher odds of report being asked for identification (OR = 6.3,95% CI 1.9-21.5). Conclusions: Retail tobacco access appears to be a significant source of cigarette access among Canadian youth. Retail tobacco access varies significantly by province, which suggests provincial policies should be strengthened. © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
Tobacco use is the leading cause of preventable death in Canada and globally (Pipe et al., 2011; Van Meijgaard & Fielding, 2012), responsible for the mortality of one in ten adults worldwide (World Health Organization, 2008). Although many interventions have shown some success in reducing smoking prevalence, a substantial proportion of youth continue to smoke (Reid, Hammond, Burkhalter, Rynard, & Ahmed, 2013). Adolescence is a critical time for smoking initiation (Levy, Friend, Holder, & Carmona, 2001), with approximately
* Corresponding author at: Propel Centre for Population Health Impact, University of Waterloo, 200 University Ave W, Waterloo, Ontario N2L2T5, Canada. E-mail address: lminaker@uwaterloo.ca (L.M. Minaker).
80% of daily smokers starting before they are 18 years old (Hammond, 2005).
To manage the burden of health costs from youth tobacco use, many countries have implemented legislation specifying a minimum age for retail tobacco access to protect adolescents. Legislation that successfully minimizes retail tobacco sales to youth directly affects youth who smoke daily by making cigarettes less available, and indirectly affects youth who smoke less frequently as daily smokers become less willing to share difficult-to-access cigarettes (DiFranza, Savageau, & Fletcher, 2009). In Canada, the Federal Tobacco Act prohibits the sale of cigarettes to anyone under the age of 18. It requires retailers to request age and photo identification of tobacco purchasers when age is in question (CRG Consulting, 2009). Nine (of 10) Canadian provinces have implemented additional tobacco control legislation, although legislation varies among the provinces (Ontario Tobacco Research Unit, 2013). Retailer compliance (i.e., the percent of retailers who do not sell tobacco
http: //dx.doi.org/10.1016/j.addbeh.2015.06.047
0306-4603/© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
to minors) has increased over time in Canada, with a national rate of 47.9% 1995 and 84.3% in 2009 (the last year for which data are available) (Health Canada, 2010a). Provincial compliance rates in 2009 ranged from 80.3% in Nova Scotia to 94.3% in British Columbia (CRG Consulting, 2009). Provincial variation in implementing youth tobacco access legislation is an important, policy-relevant topic given its potential impact on preventing tobacco-related chronic disease and the provincial legislative power to enact tobacco retail access policies (DiFranza et al., 2009).
Although the national retailer compliance rate in Canada have increased over time (Ontario Tobacco Research Unit, 2013), it falls short of the 90% to 100% compliance required to observe a real impact on youth smoking rates (Leatherdale, 2005; Rigotti et al., 1997). As of 2009, only one province, British Columbia, reached a compliance rate above 90% (CRG Consulting, 2009). In Canada, test shopping protocol to assess tobacco retailer compliance requires that hired youth (15-17 years) must dress to appear their age and truthfully answer if a store clerk asks their age (Association of Public Health Epidemiologists in Ontario, 2009). The protocol has been criticized since it does not accurately reflect the purchasing patterns of young smokers (Diemert, Dubray, Babayan, & Schwartz, 2013) and likely underestimates the number of retailers that sell tobacco to youth. Compared to standard protocol, tobacco sales to minors are significantly higher when test shoppers dress as they choose, lie about their age, purchase other items with the tobacco product and present "valid" ID if requested (DiFranza, Savageau, & Bouchard, 2001) or are familiar to the retailer (Klonoff, Landrine, & Alcaraz, 1997). Certainly, retailer compliance is different from self-reported youth retail access to tobacco. However, examining self-reported youth retail access to tobacco in light of reported retailer compliance may shed additional light on this important topic.
Age verification significantly reduces the odds of retailers selling cigarettes to minors, though it does not guarantee that illegal sales are not made (Clark, Natanblut, Schmitt, Wolters, & Iachan, 2000; DiFranza et al., 2001; Landrine & Klonoff, 2003). In the U.S., the proportion of youth who were refused a cigarette sale because they were under-aged stayed consistent between 2000 and 2009 (Filippidis, Agaku, Connolly, & Vardavas, 2014). The same national survey found that less than half of youth who reported trying to buy cigarettes in the past 30 days were asked to show proof of age over that time and more than 60% of minors who tried to buy cigarettes were successful (Filippidis et al., 2014). Clearly, youths' experiences with tobacco retailers, such as age verification, can impact youth tobacco access, which ultimately contributes to youth smoking rates (DiFranza, 2012). Youth characteristics associated with increased odds of purchasing cigarettes from retail outlets include sex, age, smoking intensity, and ethnicity. Some studies found that females are more likely to successfully purchase cigarettes from retailers (Clark, Natanblut, Schmitt, Wolters, & Iachan, 2000; Klonoff et al., 1997; O'Grady, Asbridge, & Abernathy, 1999), while others found that males are more likely to successfully buy cigarettes (Leatherdale, 2005). Older youth and heavier smokers are consistently more likely to obtain cigarettes from retail sources than younger adolescents and less heavy smokers (Health Canada, 2010a; Klonoff et al., 1997; Leatherdale, 2005; Pearson, Song, Valdez, & Angulo, 2007; Vu, Leatherdale, & Ahmed, 2011). Each year a student smokes, there is a 10% increase in the likelihood that they will buy their own cigarettes (Nelson, Paynter, & Arroll, 2011), whereas novice smokers are more likely to obtain cigarettes through friends (DiFranza et al., 2009) and family (Leatherdale, 2005; Vu et al., 2011). Additionally, youth who identify as ethnic minorities may be more likely to buy cigarettes relative to youth who identify as white (Klonoff et al., 1997; Vu et al., 2011). Finally, youth with more spending money per week are also more likely to rely on retail sources (Vu et al., 2011), and often do not spend their own money until they develop the need to smoke daily (DiFranza et al., 2009). Only one study to date has examined student- and school-level characteristics associated with youth
tobacco access in a national Canadian sample (Leatherdale, 2005). This study, using data collected in 2006/2007, found significant between-school variability in the odds of smoking students reporting buying their own cigarettes and a variety of individual-level factors associated with increased odds of purchasing cigarettes (including male gender, higher grades, Aboriginal status, higher weekly spending money, and not having family members or friends who smoke) (Leatherdale, 2005). No extant studies have examined how retail tobacco access varies by different geographic region in Canada. Examining geographic differences in youth retail tobacco access can help set the stage to determine which provincial retail tobacco restrictions are most effective in the Canadian context.
This paper uses data from the 2010/2011 national Youth Smoking Survey to describe differences in youth retail tobacco access by province. Additionally, this paper explores youths' perceptions of ease of access to tobacco and identifies youth characteristics associated with the odds of purchasing cigarettes from stores.
2. Methods
2.1. Study design
The 2010-2011 Youth Smoking Survey (2010 YSS) is a valid and reliable machine-readable, pencil and paper, nationally generalizable school-based survey that is used to measure the determinants of youth smoking behavior (Propel Centre for Population Health Impact, 2013). The target population for the 2010-2011 YSS consisted of all young Canadian residents in grades 6 to 12 attending public and private secondary schools in nine Canadian provinces (n = 426 schools). Those residing in New Brunswick, Yukon, Nunavut and Northwest Territories and those living in institutions or on First Nations reserves were not included in the target population. Participation to the 2010/2011 YSS was declined by the province of New Brunswick. Based on the comparative analysis conducted using 2008/2009 survey data, there were no statistically significant differences in national estimates with and without New Brunswick. Surveys were pilot tested to assess the logic and student understanding of the questions. Approximately three-quarters (76%) of respondents participated with passive parental permission and one-quarter (24%) participated with active parental permission. The YSS survey was administered during class time and participants were not remunerated. Survey development, design, weights, and data collection protocol for the 2008 YSS have been published (EltonMarshall et al., 2011); relevant methodological differences between the 2008/2009 and 2010/2011 YSS are highlighted below. Across Canada, 56% of schools that were invited agreed to participate. Student response rates (the percent of eligible students who completed the survey) was 73%. Preliminary analyses indicated that tobacco use among students in elementary schools was not sufficiently prevalent to provide stable estimates; therefore, data used in the following analyses were restricted to the secondary school portion of2010 YSS, which was administered to 31,396 youth in grades 9 to 12 (approximately ages 14-17). In Quebec, high school ends at grade 11, therefore grades 9-11 students from Quebec were included in this study. Data were collected between October 2010 and June 2011 and were analyzed in 2013. This study was approved by the University of Waterloo Human Research Ethics Committee, the Health Canada Research Ethics Board, and appropriate School Board and Public Health Ethics committees.
2.2. Measures and data sources
Among cigarette smokers who reported smoking in the last 30 days (N = 4523), cigarette access was determined with the question, "Where do you usually get your cigarettes?" For the dependent variable, the source of buying cigarettes was dichotomized as "usually buys from a store" (response option "I buy them myself at a store") and "usually gets from a social source" (including response options "I buy them
from a friend", "I buy them from someone else", "I ask someone to buy them for me", "My brother or sister gives them to me", "My mother or father gives them to me", "A friend gives them to me", "Someone else gives them to me", "I take them from my mother, father or siblings", "Other").
Age verification was derived from responses to the question, "In the last 30 days, have you ever been asked for ID when buying cigarettes in a store?" Response options included "I did not buy or try to buy cigarettes in a store in the last 30 days", "Yes, I was asked for ID", and "No, I was not asked for ID". This outcome was examined among students who responded either "yes" or "no".
Independent variables included the respondent's province, gender, grade (Health Canada, 2010a; Leatherdale, 2005; Ontario Tobacco Research Unit, 2013; Rigotti et al., 1997), self-reported ethnicity (White, Black, Asian, Aboriginal, Latin American, or "other"), school region urban or rural status, students' perceptions of how "easy" it is to get cigarettes if they wanted to smoke (4 levels), and the amount of weekly spending money the participant receives. Weekly spending money was categorized as $0, $1 -10, $11 -40, and > $40. Two covariates were also included in analyses: whether any family member smokes (yes/no), and the number of closest friends who smoke (none; 1; 2-4; and >4).
2.3. Statistical analysis
Survey weights were used to adjust for sample selection (school and class levels), non-response (school, class, and student levels), and poststratification of the sample population relative to grade and sex distribution in the total population. Independent variables and covariates were independently assessed as confounders to ensure that models included relevant confounding variables.
Descriptive statistics were employed to show the prevalence of grades 9-12 students who reported that it would be easy to get cigarettes by province of residence, gender, grade, self-reported ethnicity, and weekly spending money. Descriptive statistics were also employed to show the prevalence of youth who reported being asked for ID when buying cigarettes in a store in the last 30 days.
Logistic regression models were fitted to examine independent variables related to the odds of buying cigarettes from a store vs. getting cigarettes from a social source using PROC SURVEYLOGISTIC in SAS 9.3.2 (SAS/STAT software, SAS Institute Inc., Cary, NC). Logistic regression models were also fitted to examine independent variables related to the odds of being asked for identification vs. not being asked for identification among students who bought cigarettes in the previous six months. The variable was retained in the final model if it was significantly associated with the outcome (at p < 0.05) at any level of the variable. Separate models were initially constructed for each independent variable, and consisted of a main exposure (each independent variable), outcome, and potential confounders (other independent variables and covariates). Assumptions of logistic regression (e.g. sufficient sample size for single cell counts) were checked and Goodness-of-fit test were used to check model fit. A backward elimination strategy was employed in order to evaluate each covariate in the presence of others. Potential confounders were removed one at a time if change-in-estimate of effect measure was < 10%. The final models showed the measure of association between an independent variable of interest and outcome after controlling for covariates that were determined to be significant through the backward elimination strategy. Significant covariates in each model are noted in the comment below Tables 2 and 3.
3. Results
Among students who never smoked cigarettes, 78.7% of grades 9-12 students in Canada thought it would be easy to get cigarettes if they wanted to smoke (see Table 1). Among students who ever smoked cigarettes and students who smoked in the last 30 days, 93.6% and
Table 1
Weighted prevalence grade 9-12 students who thought "it is easy to get cigarettes" select characteristics - YSS 2010/2011.
Characteristics of Among never Among ever Among past 30
survey population smokers smokers day smokers
(N = 19,069) (N = 12,327) (N = 5035)
Na (weighted Na (weighted Na (weighted
prevalence (%)) prevalence (%)) prevalence (%))
Canada 19069 (78.7) 12327 (93.6) 5035 (95.8)
Gender
Female 9951 (74.6) 5756 (92.9) 2196 (96.4)
Male 9118 (83.0) 6571 (94.1) 2839 (95.3)
9 5983 (69.5) 2213 (90.1) 807 (91.8)
10 5442 (74.5) 3264 (93.3) 1275 (94.7)
11 4457 (94.7) 3602 (93.9) 1545 (96.1)
12 3187 (88.0) 3248 (95.3) 1408 (98.0)
Provinces
Atlantic 4920 (84.2) 3896 (94.5) 1760 (95.5)
Quebec 848 (85.0) 654 (96.9) 207 (98.8)
Ontario 4329 (76.3) 2083 (90.8) 764 (94.9)
Manitoba 3369 (78.4) 1856 (92.5) 687 (95.1)
Saskatchewan 806 (75.6) 802 (91.2) 386 (92.6)
Alberta 1296 (75.7) 1192 (96.1) 562 (96.2)
British Columbia 3501 (80.7) 1754(94.2) 669 (95.5)
Ethnicity
White 14109 (80.4) 9445 (94.6) 3890 (96.4)
Black 583 (84.0) 436 (92.2) 217 (92.1)
Asian 2635 (68.6) 774 (85.8) 209 (91.0)
Aboriginal 400 (71.3) 872 (95.3) 450 (96.8)
Latin American 294 (85.8) 234(89.3) 80 (99.3)
Other 892 (76.8) 487 (89.7) 155 (91.1)
Living area
Urban 14678 (79.0) 8566 (93.5) 3369 (95.4)
Rural 4391 (77.6) 3761 (93.9) 1666 (96.9)
Weekly spending money
No money 3539 (74.5) 1486 (90.6) 490 (92.0)
$1-10 2840 (69.7) 1230 (90.4) 467 (93.4)
$11-40 5495 (79.1) 3781 (92.9) 1500 (96.3)
More than $40 3839 (88.2) 3974 (96.0) 1816 (96.8)
Non-weighted distribution of variables in the whole study population.
95.8% thought it would be easy to get cigarettes, respectively. Among all categories of youth (never smokers, ever smokers and last-30-day smokers), the percent of students agreeing that it would be easy to get cigarettes was higher as grade increased. Among never smokers, the percent of students reporting that it would be easy to get cigarettes ranged by province from 75.6% in Saskatchewan to 85.0% in Quebec, and ranged by self-reported ethnicity from 68.6% of Asian never smokers to 85.8% of Latin American never smokers. Youth attending urban and rural schools reported similar rates of perceiving it would be easy to access cigarettes (79.0% in urban areas and 77.6% in rural areas).
One-quarter of last-30-day smokers reported usually buying cigarettes from a store (25.3%), whereas 74.7% reported usually getting cigarettes from a social source (see Table 2). Males had significantly higher odds of buying cigarettes from a store relative to females (OR = 1.5,95% CI: 1.1-2.2). Odds of buying cigarettes from a store were higher among grades 11 and 12 students: students in grade 11 had 3.1 times higher odds of buying cigarettes from a store relative to grade 9 students (95% CI 1.5-6.5) and students in grade 12 had 8.4 times higher odds of buying cigarettes from a store relative to grade 9 students (95% CI 3.917.9). Compared to students in Ontario, students who smoked in the last 30 days had significantly higher odds of buying cigarettes in a store in every other province except British Columbia. Last-30-day smokers in Quebec had 6.3 higher odds of buying cigarettes from a store (95% CI 3.2-12.4) as last-30-day smokers in Ontario, with 35.8% of last-30-day smokers reporting usually buying cigarettes from a store. No other independent variable significantly predicted higher or lower odds of buying cigarettes from a store.
Table 2
Factors associated with "buying cigarettes from a store" among grades 9-12 cigarette users in past 30 days (N = 4523) — YSS 2010/2011.
Predictors Weighted percent estimate Logistic regression model (1 = buying from a store, 0 = getting from a social resource)
Usually gets from a social resource (n = 3442) Usually buys from a store (n = 1081) OR unadjusted [95% CI] OR adjusted [95% CI]
Canada 74.7 25.3
Gender
Female (ref) 79.2 20.8 1.0 1.0
Male 71.4 28.6 1.4 [1.0, 2.1] 1.5 [1.1, 2.2]
9 (ref) 85.4 14.6 1.0 1.0
10 85.3 14.7 1.0 [0.5, 2.1] 1.1 [0.5, 2.6]
11 74.1 25.9 2.3 [1.1,4.5] 3.1 [1.5,6.5]
12 63.4 36.6 3.9 [1.9, 7.9] 8.4 [3.9,17.9]
Provinces
Atlantic 74.5 25.5 1.5 [1.0, 2.2] 2.2 [1.5, 3.4]
Quebec 64.2 35.8 2.2 [1.2, 3.9] 6.3 [3.2,12.4]
Ontario (ref) 80.1 19.9 1.0 1.0
Manitoba 69.9 30.1 1.8 [1.1, 2.7] 2.5 [1.5,4.1]
Saskatchewan 66.0 34.0 2.4 [1.4, 4.1] 3.8 [2.2, 6.6]
Alberta 68.7 31.3 2.6 [1.3,4.9] 3.4 [1.8, 6.5]
British Columbia 83.6 16.4 1.0 [0.6,1.8] 1.3 [0.7,2.5]
Ethnicity
White (ref) 74.2 25.8 1.0 1.0
Black 77.4 22.6 1.1 [0.4, 2.6] 0.9 [0.3, 2.9]
Asian 78.4 21.6 0.7 [0.3, 2.0] 0.8 [0.2, 2.8]
Aboriginal 75.9 24.1 1.0 [0.6,1.7] 1.2 [0.7, 2.1]
Latin America/Hisp 86.8 13.2 1.0 [0.3,2.7] 1.3 [0.5,3.2]
Other 64.4 35.6 1.6 [0.7, 3.9] 2.3 [0.7, 7.6]
Living area
Urban (ref) 74.4 25.6 1.0 1.0
Rural 75.9 24.1 0.8 [0.6,1.2] 0.8 [0.5,1.2]
Weekly spending money
No money 78.8 21.1 1.0 1.0
$1-10 84.8 15.2 0.5 [0.2,1.3] 0.5 [0.2,1.5]
$11-40 77.7 22.3 0.8 [0.4,1.6] 0.8 [0.4,1.7]
More than $40 68.2 31.8 1.4 [0.8, 2.6] 1.1 [0.5, 2.2]
Thought that it is easy to get cigarettes
No 75.7 24.3 1.0 1.0
Yes 73.3 26.7 1.2 [0.5, 2.8] 0.6 [0.2,1.8]
For gender and province, the model included grade, gender, province, and living area. For grade, the model is adjusted for gender, province, and number of closest friends smoking. For ethnicity, living area, and weekly spending money, the models are adjusted for gender, grade, province, and number of closest friends smoking.
Students in grade 12 had significantly higher odds of being asked for identification when buying cigarettes in a store than grade 9 students (OR = 6.3, 95% CI 1.9-21.5) (see Table 3). No other variables were significantly associated with the odds of students reporting being asked for identification when they bought cigarettes in a store.
4. Discussion
In this nationally generalizable sample, almost 80% of Canadian grades 9-12 never smokers and 96% of current smokers thought it would be easy to get cigarettes. This study contributes two major findings to knowledge about youth tobacco use in Canada. First, nationally, about one-quarter of Canadian youth smokers reported usually buying cigarettes from a store. Importantly, however, provinces varied significantly in terms of the prevalence of smokers usually buying cigarettes from stores from a low of 16% in British Columbia to a high of 36% in Quebec. This finding is particularly of interest given the national retailer compliance estimate, as will be discussed below. Second, grade 12 students had significantly higher odds of buying cigarettes from a store and also significantly higher odds of being asked for ID when buying cigarettes relative to grade 9 students. Each of these findings is described in more detail below.
The first major contribution of this study is finding that youth retail tobacco access varies significantly by province. The percent of smokers who usually bought cigarettes from a store ranged from a low of 16% in British Columbia to a high of over double that (36%) in Quebec. This variation indicates a significant role for policy to narrow provincial
gaps in youth retail tobacco access. The self-reported youth retail tobacco access we examined in this study is of particular interest when juxtaposed with national retailer compliance data. In 2009, the last year Health Canada conducted a national tobacco retailer compliance evaluation, 84% of tested retailers (including chain convenience stores, grocery stores, gas stations, and independent convenience stores) refused to sell cigarettes to "test shoppers" (hired 15-17 year-olds who attempt to purchase cigarettes). These rates remained steady from 2007-2009 and reflect increases in retailer compliance since 2003. Econometric analyses indicate that improved rates of retailer compliance help account for reduced rates of youth smoking in Canada (Health Canada, 2010b). Fig. 1 shows reported retailer compliance from 2009, as per Health Canada's test shopper protocol (described above) and the percent of smokers who reported usually buying cigarettes from a store from YSS 2010/2011 data. Retailer compliance rates from Health Canada do not appear to mirror self-reported youth retail tobacco access from our study, although British Columbia (the only province to have a retailer compliance rate over 90%) has both the highest retailer compliance and the lowest percent of student smokers who usually buy their cigarettes at stores. Perhaps this reflects previous findings that retailer compliance needs to be at least 90% to be effective in disrupting youth retail tobacco access (Leatherdale, 2005; Rigotti et al., 1997). Future research should examine reasons for discrepancies in youth retail tobacco access between provinces. An evaluation of U.S. state policies found no association between the strength of a law and the strength of associated enforcement (DiFranza et al., 2009), indicating that future evaluations of Canadian provincial
Factors associated with "being suggested a particular brand" (N = 1187) and "being asked for ID in the last 30 days" (N = 1006) among grades 9-12 students who reported buying cigarettes in a store in the last 6 months — YSS 2010/2011.
Predictors Weighted percent estimate Weighted percent estimate Logistic regression model (1 =
asked for ID, 0 = not asked for ID)
Store clerk did not suggest a brand Store clerk suggested a brand Was not asked for ID (n Asked for ID (n = OR unadjusted OR adjusted
(n = 984) (n = 203) = 523) 483) [95% CI] [95% CI]
Canada 82.5 17.5 50.9 49.1
Gender
Female (ref) 88.2 11.8 50.1 49.9 1.0 1.0
Male 79.5 20.5 51.3 48.7 1.0 [0.5,1.8] 1.0 [0.5,1.9]
9 (ref) 62.8 37.2 79.2 20.8 1.0 1.0
10 77.1 22.9 53.8 46.2 3.0 [0.7,12.1] 2.7 [0.7,11.1]
11 79.2 20.8 50.3 49.7 3.2 [0.9,11.7] 3.4 [1.0,11.9]
12 88.8 11.2 45.8 54.2 3.9 [1.1,14.1] 6.3 [1.9,21.5]
Provinces
Ontario (ref) 78.6 21.4 51.0 49.0 1.0 1.0
Atlantic 84.0 16.0 61.7 38.3 0.6 [0.3,1.1] 0.7 [0.4,1.3]
Quebec 91.7 8.3 47.7 52.3 1.2 [0.5, 2.9] 2.4 [1.0, 5.9]
Manitoba 87.6 12.4 36.7 63.3 1.8 [0.9, 3.6] 2.1 [1.1,4.3]
Saskatchewan 85.7 14.3 34.6 65.4 1.9 [0.8,4.1] 2.1 [1.0,4.6]
Alberta 80.8 19.2 60.7 39.3 0.6 [0.2,1.7] 0.6 [0.2,1.7]
British Columbia 72.1 27.9 51.7 48.3 0.9 [0.4, 2.5] 1.1 [0.4, 3.2]
Ethnicity b
White (ref) 84.5 15.5 49.8 50.2 1.0 1.0
Black 60.5 39.5 77.1 22.9 0.3 [0.1,1.1] 0.3 [0.1,1.1]
Asian 83.5 16.5 46.4 53.6 1.1 [0.3,4.5] 1.3 [0.3,5.1]
Aboriginal 92.4 7.6 47.1 52.9 1.1 [0.5, 2.3] 0.8 [0.4,1.6]
Latin America/Hisp 88.7 11.3 42.8 57.2 1.5 [0.3, 7.8] 1.6 [0.2,12.0]
Other 49.3 50.7 58.8 41.2 0.7 [0.2, 2.3] 1.0 [0.3,3.3]
Living area
Urban (ref) 81.9 18.1 51.7 48.3 1.0 1.0
Rural 84.8 15.2 48.2 51.8 1.1 [0.6, 2.0] 1.1 [0.6, 2.1]
For gender, grade, provinces, ethnicity, and living area, the logistic regression model includes all other independent variables and covariates.
legislation must consider both the law as well as enforcement in defining "policy strength".
The second major contribution of this study is the finding that compared to older students, younger students had lower odds of reporting being asked for ID. Importantly, our findings that the percent of youth smokers who usually buy cigarettes from a store is higher among students in higher grades reflects findings from the national retailer compliance study showing rates of refusal were 95% for 15 year olds, 83% for 16 year olds and 80% for 17 year olds (CRG Consulting, 2009; Health Canada, 2010a). However, findings from Health Canada's 2009 retailer behavior study suggest 83% of retailers asked test shoppers for
ID (CRG Consulting, 2009). The high prevalence of retailers asking for ID is not reflected in our study: only 49% of students who attempted to purchase cigarettes in the last 30 days reported being asked for ID (which is comparable to self-reported age verification prevalence among American youth) (Filippidis et al., 2014). Furthermore, about one in five grade 9 students (generally aged 13-14 years) who bought cigarettes from a store reported being asked for ID in the last 30 days, compared to over half (54%) of grade 12 students (generally aged 1718 years). This finding was somewhat counter-intuitive, given that retailers are expected to ask for ID when customers appear to be younger than the purchase age (18 or 19 years; purchase age varies by province).
100 90 80 70 60 50 40 30 20 10 0
I Retailer Compliance (Health Canada, 2009)
1% smokers in grades 9-12 who usually buy cigarettes at a store (YSS, 2010/2011)
Fig. 1. Health Canada's retailer compliance rates (2009) and the percent of grades 9-12 smokers who usually buy cigarettes at a store (YSS 2010/2011).
It may be that students in grade 9 were denied a sale before being asked for ID. Alternatively, it is possible that grade 9 students only go to stores where they know the clerks will sell them cigarettes and thus are not asked for ID, whereas grade 12 students might be more likely to attempt to buy cigarettes at stores where they are not "known" to the clerk. Findings from our study suggest that additional modifications to current test shopping protocols may better reflect adolescent purchasing patterns. For example, including test shoppers younger than 15 years might reveal different patterns of retailers' refusal, since in our study, grade nine students (ages 13-14) had lower odds of reporting that they were asked for ID in the last 30 days relative to grade 12 students.
Despite this study using a large, nationally and provincially-representative sample of students in grades 9-12, there are several limitations. Given the cross-sectional nature of the survey, causation cannot be inferred. In addition, the YSS is a school-based survey. Therefore, youth who do not attend eligible schools or are not enrolled in school were excluded from our sample, as were Aboriginal students living on reserve and youth living in Canada's three territories. Third, the YSS 2010/2011 did not contain questions on the frequency of attempts to buy cigarettes from a retailer. Fourth, the data are self-reported, although any potential misclassification is expected to be non-differential. Finally, we did not examine provincial-level characteristics such as retail tobacco access policy existence or implementation.
In terms of policy relevance, this study found that despite high levels of reported retailer compliance (CRG Consulting, 2009), one-quarter of Canadian smokers in grades 9-12 report usually buying their cigarettes at a store. Future research should examine the strength of the provincial legislation and implementation and its relationship with youth retail tobacco access and youth smoking to determine the most effective and cost-efficient components of existing legislation and regulations.
Competing interests
The authors have no competing interests.
Funding
This work was supported by the Canadian Cancer Society major program grant number 701019. The Youth Smoking Survey is supported by Health Canada contract number H4133-122931-001-SS.
Contributorship
LMM wrote the manuscript, contributed to the analyses and incorporated co-authors' edits and feedbacks into the final manuscript version. SS co-wrote the manuscript. NN conducted the analyses. SM contributed to the manuscript writing and provided substantial feedback on drafts. All of the authors listed have agreed to submission of the manuscript in this form. LMM has assumed responsibility for keeping the coauthors informed of this manuscript's progress through the editorial review process, the content of the reviews, and any revisions made.
Acknowledgments
The Youth Smoking Survey is a product of the pan-Canadian capacity building project funded through a contract between Health Canada and the Propel Centre for Population Health Impact from 2008 through 2011. The YSS consortium includes Canadian tobacco control researchers from all provinces and provided training opportunities for university students at all levels. The views expressed herein do not necessarily represent the views of Health Canada. The Propel Centre for Population Health Impact acknowledges the support of the Canadian Cancer Society [Major Program Grant #701019].
References
Association of Public Health Epidemiologists in Ontario (2009). 5Aminors' access to tobacco.
Available at: http://www.apheo.ca/index.php?pid=120 (Accessed 07/25,2014).
Clark, P.I., Natanblut, S.L., Schmitt, C.L., Wolters, C., & Iachan, R. (2000a). Factors associated with tobacco sales to minors. Lessons learned from the FDA compliance checks. Journal of the American Medical Association, 284(6), 729-734.
Clark, P.I., Natanblut, S.L., Schmitt, C.L., Wolters, C., & Iachan, R. (2000b). Factors associated with tobacco sales to minors. Lessons learned from the FDA compliance checks. Journal of the American Medical Association, 284(6), 729-734.
CRG Consulting (2009). Evaluation of retailers' behaviour towards certain youth access-to-tobacco restrictions: final report findings 2009. (H4133-081900/001 /SS).
Diemert, L., Dubray, J., Babayan, A., & Schwartz, R. (2013). Strategies affecting tobacco vendor compliance with youth access laws. Available at: http://otru.org/wp-content/uploads/2013/10/special_vendor_compliance.pdf (Accessed 07/24,2014).
DiFranza, J.R. (2012). Which interventions against the sale of tobacco to minors can be expected to reduce smoking? Tobacco Control, 21(4), 436-442.
DiFranza, J.R., Savageau, J.A., & Bouchard, J. (2001). Is the standard compliance check protocol a valid measure of the accessibility of tobacco to underage smokers? Tobacco Control, 10(3), 227-232.
DiFranza, J.R., Savageau, J.A., & Fletcher, K.E. (2009). Enforcement of underage sales laws as a predictor of daily smoking among adolescents — a national study. BMC Public Health, 9.
Elton-Marshall, T., Leatherdale, S.T., Manske, S.R., Wong, K., Ahmed, R., & Burkhalter, R. (2011). Research methods of the Youth Smoking Survey (YSS). Chronic Diseases and Injuries in Canada, 32(1), 47-54.
Filippidis, F.T., Agaku, I.T., Connolly, G.N., & Vardavas, C.I. (2014). Trends in age verification among U.S. adolescents attempting to buy cigarettes at retail stores, 2000-2009. Preventive Medicine, 61, 61-65.
Hammond, D. (2005). Smoking behaviour among young adults: beyond youth prevention. Tobacco Control, 14,181-185.
Health Canada (2010a). Retailer compliance evaluation. Available at: http://www.hc-sc.gc. ca/hc-ps/tobac-tabac/research-recherche/eval/index-eng.php (Accessed 07/25,2014).
Health Canada (2010b). Evaluation of retailers' behaviour towards certain youth access-to-tobacco restrictions. Available at: http://www.hc-sc.gc.ca/hc-ps/tobac-tabac/ research-recherche/eval/2009_result-eng.php (Accessed 07/28, 2014).
Klonoff, EA, Landrine, H., & Alcaraz, R. (1997). An experimental analysis of sociocultural variables in sales of cigarettes to minors. American Journal of Public Health, 87(5), 823-826.
Landrine, H., & Klonoff, EA. (2003). Validity of assessments of youth access to tobacco: the familiarity effect. American Journal of Public Health, 93(11), 1883-1886.
Leatherdale, S.T. (2005). Predictors of different cigarette access behaviours among occasional and regular smoking youth. Canadian Journal of Public Health, 96(5), 348-352.
Levy, D.T., Friend, K., Holder, H., & Carmona, M. (2001). Effect of policies directed at youth access to smoking: results from the SimSmoke computer simulation model. Tobacco Control, 10(2), 108-116.
Nelson, R., Paynter, J., & Arroll, B. (2011). Factors influencing cigarette access behaviour among 14-15-year-olds in New Zealand: a cross-sectional study. Journal ofPrimary Health Care, 3(2), 114-122.
O'Grady, B., Asbridge, M., & Abernathy, T. (1999). Analysis of factors related to illegal tobacco sales to young people in Ontario. Tobacco Control, 8(3), 301-305.
Ontario Tobacco Research Unit (2013). Youth access to tobacco products: monitoring update. Available at: http://nbatc.ca/en/uploads/file/Youth%20Access%20to%20Tobacco% 20Products%20_OTRU%20Sept%202013.pdf (Accessed 07/24,2014).
Pearson, D.C., Song, L., Valdez, R.B., & Angulo, A.S. (2007). Youth tobacco sales in a metropolitan county. Factors associated with compliance. American Journal of Preventive Medicine, 33(2), 91-97.
Pipe, A.L., Eisenberg, M.J., Gupta, A., Reid, R.D., Suskin, N.G., & Stone, JA. (2011). Smoking cessation and the cardiovascular specialist: Canadian cardiovascular society position paper. Canadian Journal of Cardiology, 27(2), 132-137.
Propel Centre for Population Health Impact (2013). Youth Smoking Survey 2012/2013. Available at: http://www.yss.uwaterloo.ca/index.cfm?section=1001&page=248 (Accessed 08/15,2013).
Reid, J.L., Hammond, D., Burkhalter, R., Rynard, V.L., & Ahmed, R. (2013). Tobacco use in Canada: patterns and trends (2013 ed.).
Rigotti, N.A., DiFranza, J.R., Chang, Y., Tisdale, T., Kemp, B., & Singer, D.E. (1997). The effect of enforcing tobacco-sales laws on adolescents' access to tobacco and smoking behavior. New England Journal of Medicine, 337(15), 1044-1051.
Van Meijgaard, J., & Fielding, J.E. (2012). Estimating benefits of past, current, and future reductions in smoking rates using a comprehensive model with competing causes of death. Preventing Chronic Disease, 9(7).
Vu, M., Leatherdale, S.T., & Ahmed, R. (2011). Examining correlates of different cigarette access behaviours among Canadian youth: data from the Canadian Youth Smoking Survey (2006). Addictive Behaviors, 36(12), 1313-1316.
World Health Organization (2008). The global tobacco crisis. Available at: http://www.who. int/tobacco/mpower/mpower_report_tobacco_crisis_2008.pdf (Accessed 07/24, 2014).
