Understanding Youth Trajectories of Alcohol Use: Evaluating the Impact of a Policy of Authorizing Alcohol Sales in Some Grocery Stores in Ontario

Abstract

The latest change in the policy of the Liquor Control Board of Ontario (LCBO), enacted in December 2015, authorizes the sale of alcohol in 450 grocery stores that previously had not sold alcohol. This policy change increased the availability of alcohol to the province’s population in terms of the number of off-premise alcohol outlets, hours and days of sale. Since alcohol consumption is highly prevalent among Canadian youth, the new LCBO policy may result in a change in youth alcohol use patterns. The youth population is heterogeneous with respect to alcohol use patterns, and there exist several distinct patterns of alcohol use among youth each representing a subgroup of the population. As these subgroups differ in patterns of use, their responses to changes in alcohol-related policies could likewise differ. Given the numerous immediate and long-term adverse health outcomes associated with alcohol consumption, evaluation of the impact of the new LCBO policy on each subgroup of the population is a public health priority. The research presented in this dissertation research used longitudinal data to examine the impact of the new LCBO policy on youth alcohol use patterns in a quasi-experimental setting with two control groups, one in Ontario and the other in Alberta. The objectives were to (1) identify patterns of alcohol consumption among the youth population, (2) investigate differences in alcohol use patterns across secondary schools, (3) estimate the probabilities of transitioning between drinking patterns over time, (4) examine how individual- and school-level covariates predict patterns of alcohol use and transitioning between the patterns, and (5) evaluate any impact of the new LCBO policy on changes in the alcohol use patterns of youth in the exposed jurisdictions by means of pre- and post-policy assessment. Three manuscripts served these objectives all using student- and school-level data from the COMPASS study. The first manuscript identified latent classes (patterns) of youth alcohol consumption and investigated whether the size of latent classes vary by school. The multilevel latent class analysis identified four student-level latent classes and two school-level latent classes. Student-level classes of youth alcohol use were characterized as non-drinkers (had never consumed alcohol or did not drink in past year, 44.2% of the sample), light drinkers (initiated alcohol consumption at age 14 or later and consumed alcohol up to 3 times a month, with one time binge drinking, 41.8% of the sample), regular drinkers (initiated drinking before age 14 and consumed alcohol 1 to 3 times a week and undertook binge drinking 2 to 4 times a month,11.1% of the sample), and heavy drinkers (initiated drinking before age 14 and consumed alcohol on a nearly everyday basis, with binge drinking more than once a week, 2.9% of the sample). Two groups of schools were characterized, one as low-use [schools with a relatively large number of non-drinkers (n=40, 44.9% of schools)] and the other as high-use [schools with higher rates of regular and heavy drinkers (n=49, 55.1%)]. Male students (OR=1.30) and upper grades (OR=1.93) were significantly associated with membership in higher use groups. The number of off-premise alcohol outlets and the median household income was not significantly associated with membership in school-level latent classes. The second manuscript estimated the probabilities of maintenance or change between the identified drinking patterns two years before the LCBO policy change. Latent transition analysis of longitudinal data identified a similar number of latent classes with slightly different structures. LTA identified four latent classes: non-drinker, periodic drinker (reported drinking up to 3 times a month and no binge drinking), lower risk drinker (reported drinking up to 3 times a month and once or less than a month binge drinking), and high-risk regular drinker (reported drinking 1-3 times a week and binge drinking 2-4 times a month). The results of the second manuscript suggest that alcohol consumption tends to increase among those who start, but the volume of increase is not the same across all populations of youth. The greatest risk of increasing consumption was observed among the periodic drinker group: 40% of periodic drinkers demonstrated not only an increase in their consumption but also reported occasions of binge drinking at the follow-up year, when previously they had not engaged in binge drinking at all. Overall, the proportion of individuals engaged in some level of binge drinking rose from 21.2% at baseline to 33.9% at follow-up. The manuscript suggests that some youth do quit drinking or reduce the level of their consumption over time. Nearly 10% of low risk or high-risk regular drinkers were likely to have stopped drinking one year later at the follow-up. The results indicate that the use of either cigarettes or cannabis is markedly associated with greater risk of membership in higher risk drinking classes. The third manuscript examined changes in probabilities of transitioning between the four identified patterns of alcohol use among subpopulations in the exposed jurisdiction, using the two control cohorts (respectively in Ontario and Alberta) before and after the new LCBO policy. The transition probabilities estimated by latent transition analysis suggest that the new LCBO policy has no negative impact on the periodic and low-risk drinker groups that comprise 57% of the population. In contrast, there is a negative impact among non-drinker and high-risk regular drinker groups as the lowest and highest risk groups, respectively. In communities with alcohol available in grocery stores, the lowest risk group (non-drinker) were more likely to transition to the highest risk group (high-risk regular drinker), and individuals in the highest risk group indicated a higher likelihood of maintaining their behaviour. Comparison of the prevalence of high-risk regular drinking among the three cohorts indicated that youth in the exposed jurisdictions had the highest rate of increase during the follow-up years. This dissertation enhances our knowledge of alcohol use patterns in Canadian youth, the developmental stages of alcohol use over time, and the impact of the new LCBO policy on alcohol use patterns of the youth population. This work signals the need for continued efforts to delay alcohol initiation of abstainers as the primary prevention and to promote the transition from drinking to reducing or quitting among drinkers. This dissertation illustrates that following the increase in physical availability of alcohol in the jurisdictions exposed to the new LCBO policy, the probability of high-risk regular drinking among youth has increased. Abstainers in the exposed jurisdictions had a higher risk of being regularly engaged in high-risk drinking compared to youth in unexposed jurisdictions. Further, high-risk regular drinkers are more likely to maintain their behaviours following the policy change. This research provides scientific, documentary and timely evidence on the impact of the new alcohol policy on specific subgroups of the youth population separately. This evidence can fuel and support initiatives to reduce harm of youth alcohol consumption. Moreover, the analytic approach used offers a template for future evaluations of the impacts of public policy changes, for instance, to examine patterns of cannabis use in view of federal Cannabis legalization.