The association between single and dual use of cannabis and alcohol and driving under the influence and riding with an impaired driver in a large sample of Canadian adolescents

Abstract

Objective

Dual use of cannabis and alcohol has increased in adolescents, but limited research has examined how it relates to impaired driving or riding with an impaired driver (IDR) compared to single substance use. This study aimed to examine the odds of alcohol- and/or cannabis-IDR among adolescents based on their use of alcohol and/or cannabis, and whether associations differed by gender and age.

Methods

Cross-sectional survey data were used from a sample of 69,621 students attending 182 Canadian secondary schools in the 2021/22 school year. Multilevel logistic regression estimated the odds of exclusive alcohol-IDR, exclusive cannabis-IDR, and both alcohol and cannabis IDR (alcohol-cannabis-IDR). Substance use interactions with gender and age were tested.

Results

Overall, 14.7% of participants reported IDR; 7.5% reported exclusive alcohol-IDR, 3.2% reported exclusive cannabis-IDR, 4.0% reported alcohol-cannabis-IDR, and 7.4% were unsure if they had experienced IDR. The prevalence of IDR varied across substance use groups, 8.0% among nonuse, 21.9% among alcohol-only use, 35.9% among cannabis-only use, and 49.6% among dual use groups. Gender diverse, older, and students with lower socioeconomic status exhibited a higher likelihood of reporting alcohol-cannabis-IDR. Dual use was significantly associated with 9.5 times higher odds of alcohol-cannabis-IDR compared to alcohol-only use, and 3.0 times higher odds compared to cannabis-only use. Dual use was also associated with an increased likelihood of either alcohol- or cannabis-IDR.

Conclusions

This study highlights that all students, regardless of substance use, are at risk of IDR, but students engaged in dual use of alcohol and cannabis face an elevated risk compared to both peers who do not use substances and those who use only a single substance. These findings emphasize the importance of targeted interventions that address the risks associated with IDR.

Keywords:

• Secondary school students
• substance use
• road safety
• impaired driving

Background

Adolescence, spanning ages 12–19, is a period characterized by increased risk-taking behavior, during which substance use often begins and many individuals obtain driving licenses. Vehicles may become their primary mode of transportation and provide private spaces for substance use (Hasan et al. 2022). The combination of driving and substance use elevates the risk to individual health, particularly increasing the likelihood of motor vehicle crashes. Alcohol and cannabis can contribute to risky driving behaviors such as speeding, tailgating, and abrupt stops, as well as impaired reaction time and diminished ability to divide attention between multiple tasks, which are linked to vehicle accidents (Ramaekers et al. 2011; Hartman and Huestis 2013; Pearlson et al. 2021). Vehicle accidents are a leading cause of death among adolescents (Mayhew et al. 2004; Canada PHAo 2022). According to data from U.S., 3,058 teenagers lost their lives in car accidents in 2021, meaning approximately nine teens died due to motor vehicle crashes every day, and of them, one-third had been drinking. Almost one in ten grade 11–12 students in Canada report driving within an hour of alcohol consumption (9.1%) or within 2 h of cannabis use within the past 30 days (9.4%) (Minaker et al. 2017). Riding with an impaired driver is also a significant concern. Canadian data indicate that 35% of grade 9 to 12 students report riding with an alcohol-impaired driver, and 20% report riding with a cannabis-impaired driver within past 30 days (Minaker et al. 2017).

Dual use of both alcohol and cannabis has been on the rise among adolescents and is associated with higher risk of engaging in harmful behaviors compared to single substance use (Yurasek et al. 2017; Morris et al. 2018). While the association between risky driving and single use of alcohol and cannabis has been acknowledged, limited research has explored links with dual use. Existing studies that have examined the effects of multiple substance use have predominantly focused on simultaneous use (i.e., use of more than one substances on the same occasion) rather than dual use (i.e., use of both substances but not necessarily on the same occasion) (Terry-McElrath et al. 2014; Subbaraman and Kerr 2015; Shults et al. 2019). While the increased risk associated with simultaneous use is well-documented, dual use remains a concern; research among college students has demonstrated similar odds of impaired driving among those who reported simultaneous and dual use (Jackson et al. 2020). Additionally, there is limited evidence on the specific effects of alcohol-only or cannabis-only use and the likelihood of reporting riding with an impaired driver while under the influence of the other substance. Therefore, it is important to investigate the experiences of impaired driving or riding with an impaired driver (IDR) among adolescents who engage in alcohol-only or cannabis-only use, as well as those who use both substances in order to gain a comprehensive understanding of the associated risks.

Prior research has suggested that during adolescence, IDR results from a complex interplay of various individual and social factors, some exerting direct influences and others indirect. Social factors, including the highest level of parental education, family affluence, parental monitoring knowledge, and peers’ perception of IDR, have shown significant associations with IDR (Chen et al. 2008; Li et al. 2014; Vaca et al. 2021). Additionally, individual factors such as age and gender have been identified as significant correlates of IDR (Minaker et al. 2017), along with variations in levels and patterns of alcohol and cannabis use.

From a harm reduction perspective, it is crucial to identify population subgroups that are at a higher risk for IDR. This knowledge can contribute to the development of more targeted and effective prevention strategies, taking into account the unique risk patterns associated with substance use. Additionally, the availability of both alcohol and cannabis is an important factor that may influence the prevalence of dual alcohol and cannabis use. The legalization of adult recreational cannabis use in Canada in October 2018 and the significant alcohol sale expansion in some jurisdictions, including Ontario as the most populated province of Canada, represents a significant environmental shift that may impact the prevalence of dual use and its consequences such as impaired driving. As these policy changes alter the accessibility and societal norms surrounding both cannabis and alcohol, adolescents may experience increased exposure to these substances, potentially leading to higher rates of dual use and related negative outcomes.

This study aimed to address evidence gaps by examining dual and single alcohol and cannabis use in relation to separate outcomes of IDR related to alcohol, cannabis, or both alcohol and cannabis in a large sample of Canadian secondary school students. We hypothesized that dual use of alcohol and cannabis would be associated with higher odds of alcohol and cannabis IDR compared to no use or single use of either substance. According to past research, we hypothesized that boys would be more likely to report IDR compared to girls. Given evidence of increased substance use during adolescence, and greater opportunity for driving with advancing age, we hypothesized that older secondary school students would be more likely to report IDR.

Methods

Design and participants

This research used data from the ongoing COMPASS (Cannabis use, Obesity, Mental health, Physical activity, Alcohol use, Smoking, and Sedentary behavior) study, which collects annual survey data from Canadian secondary school students (Leatherdale et al. 2014). The current study used cross-sectional data from 69,621 students between 12 and 19 years of age, attending a convenience sample of 182 schools (n = 113 Quebec, n = 51 Ontario, n = 13 British Columbia, n = 5 Alberta) that participated in COMPASS during the 2021/22 school year. Participating schools employed an active-information passive-consent parental permission protocol. In this protocol, all students in the schools were considered eligible unless their parents chose to exclude them. The average rate of parent permission in this year was 99.6%, and the participation rate was 68.5%. Data were collected online using the Qualtrics XM survey software. A survey link was emailed to all students who were not excluded from the study by their parents. Schools allocated class time for students to complete the survey; however, students were also allowed to complete the survey within a week outside of the class (Reel et al. 2021).

Measures

Impaired driving or riding with an impaired driver (IDR)

To assess alcohol-IDR, students were asked, “In the last 30 days, how many times have you done the following? Rode in/on a vehicle when the driver (you or someone else) had been drinking alcohol?” Response options included “Never,” “Once,” “Twice,” “Three or more times,” and “I don’t know.” The same question was asked for cannabis-IDR. These two measures do not differentiate between driving under the influence and riding with an impaired driver. This approach was adopted to help reduce potential bias if respondents felt concerned about implicating themselves in potentially criminal activities. As a result, the measure prevents us from examining the specific contribution of impaired drivers versus passengers with impaired drivers. Based on the frequency of alcohol- and cannabis-IDR, participants were classified to one of the five following categories: none (if responses to both alcohol and cannabis IDR questions were never), exclusive alcohol-IDR (if students reported once or more for alcohol-IDR and never for cannabis), exclusive cannabis-IDR (if students reported once or more for cannabis-IDR and never for alcohol), and both alcohol-cannabis-IDR (if students reported once or more for both alcohol and cannabis IDR). If participants reported “I don’t know” for either alcohol- or cannabis-IDR, then participants were classified in the Unsure category. The frequency of the latter group was only reported descriptively.

Substance use

Consistent with other studies (Burnett et al. 2023), and to be comparable to the time frame of the IDR measure, alcohol and cannabis use were categorized as distinct binary variables. If a student reported using the substance at least once a month in the past year, the respective binary variable was coded as one; otherwise, if there was no use or usage was less than once a month in the last year, the binary variables were coded as zero. These binary variables were then used to categorize substance use into four levels: (1) none, (2) cannabis-only, (3) alcohol-only, and (4) dual use. If a participant had missing data for either alcohol or cannabis, but provided a valid response for another, they were classified as engaging in use based on the available valid response.

Student-level covariates

Participants reported on their ethnicity (Black, East Asian, Latino, Middle Eastern/South Asian, White, mixed/Another ethnicity), gender (cis-gender girl, cis-gender boy, and gender diverse), and age. Gender was determined based on a two-step measure assessing sex assigned at birth and gender identity (Lagos and Compton 2021). To account for the age of legal driving, which is typically 16 years across all jurisdictions in our study, age was included as a binary variable distinguishing between ages 15 and under, and ages 16 and above. Socio-economic status was measured using two proxy questions. The first question assessed student-level income by asking students how much money they usually get to spend or save weekly with response groups of zero, $1–$20, $21–$100, >$100, and I don’t know. The second question involved a subjective rating of relative socio-economic position, where students indicated whether they perceived their family to be more, less, or as financially comfortable compared to other families in their class. These two variables were included in the models as distinct categorical variables.

School-level characteristics

School-level characteristics were collected in relation to urbanicity (rural/small, medium, large), and province (Alberta, British Columbia, Ontario, and Quebec). Urbanicity was obtained from the 2021 census Canada data (Canada 2023). Urban and rural categories were based on Statistics Canada’s Statistical Area Classification system and were derived from school postal codes. Large urban areas were considered census divisions with populations more than 100,000 and medium urban areas were census divisions with a population between 30,000 to 99,999. All other areas were considered rural/small urban.

Data analysis

We initially determined the proportion of students who reported exclusive alcohol-IDR, exclusive cannabis-IDR, and alcohol-cannabis-IDR. Separate univariate logistic regression analyses with only one predictor were performed to explore the associations between potential individual-level covariates (i.e., gender, age, ethnicity, weekly spending money, family affluence) and school-level covariates (i.e., province and urbanicity) and the three types of IDR (i.e., exclusive alcohol-IDR, exclusive cannabis-IDR, and alcohol-cannabis-IDR). Subsequently, significant covariates identified in the univariate models were incorporated into the multivariable logistic regression analysis, along with substance use variables. These models aimed to examine the relationships between substance use and experiencing exclusive alcohol-IDR, exclusive cannabis-IDR, and alcohol-cannabis-IDR, while accounting for significant covariates at the univariate levels.

To address the confounding effects of including participants who do not use substances, we performed another analysis specifically focusing on participants who reported substance use. The analysis included two separate sets of multilevel binary regressions. In these models we estimated the risk of IDR among dual use and single use groups. The first set estimated the odds of the three IDR outcomes for dual use compared to alcohol-only use, while the second set estimated the odds for dual use compared to cannabis-only use. All models adjusted for significant variables identified in the univariate logistic regressions, province, and accounted for the nested structure of the data, with students’ responses nested within schools.

To explore potential moderating effects of gender and age, interaction terms between substance use and gender, as well as substance use and age, were included in the models. All statistical analyses were performed using SAS Studio (Enterprise Edition).

Missing values

We excluded students who did not respond to any of the IDR questions (n = 5,521) from the original sample. Among the remaining 69,621 students, the proportion of students with complete data was 96.3% (n = 67,011), and the remaining 3.7% (n = 2,610) students had missing data on one or multiple variables. Of the students with missing data, 2,352 had only one, 189 had two, and 69 had more than two missing values. The rates of missing alcohol and cannabis IDR were 0.08% (n = 55) and 0.5% (n = 362), respectively. Substance use data was missing for 0.9% (n = 624) of students. Ethnicity had the highest missing rate of the variables (2.1%). Because of low missing rates, we did not use any imputation method for missing data. All covariates were statistically significant in the univariate regressions and were consequently included in the final model. The final analysis sample for the multivariable regression models consisted of individuals with no missing covariates (n = 67,011).

Results

The average age of the study participants was 14.8 (SD = 1.5), with 33.8% (n = 23,516) aged 16 or older (Supplementary Table A1). The mean ages were relatively consistent across gender groups, with girls averaging 14.8, boys 14.7, and gender-diverse students 14.9. A predominant portion of participants self-identified as White (70.6%) and described their family affluence as comfortable compared to other families in the community (66.7%). Half of the participants attended schools located in large cities (>100,000 population). The majority of participants (66.4%, n = 45,813) reported no substance use, while 24.7% (n = 17,059) reported alcohol-only use, 2.2% (n = 1,488) reported cannabis-only use, and 6.7% (n = 4,637) reported dual use. Overall, 14.7% (n = 10,294) of participants engaged in either alcohol-IDR or cannabis-IDR within the past 30 days (Figure 1). Specifically, 7.5% (n = 5,245) experienced exclusive alcohol-IDR, 3.2% (n = 2,246) experienced exclusive cannabis-IDR, and 4.0% (n = 2,803) experienced alcohol-cannabis-IDR. Notably, 7.4% (n = 5,178) of students reported being unsure if they had experienced either alcohol- or cannabis-IDR, with 5.9% (n = 4,113) unsure about alcohol-IDR and 4.2% (n = 2,887) unsure about cannabis-IDR. The association between substance use and IDR is depicted in Figure 2. Among adolescents who reported no substance use, 8.0% (n = 3,669) reported IDR, which is significantly lower compared to those who reported substance use. Overall, the frequency of IDR was higher among the alcohol-only use group (n = 3,738) compared to the cannabis-only use group (n = 531) and dual use group (n = 2,292). However, the proportion of adolescents in the cannabis-only use group who reported IDR (35.9%) was higher than that of adolescents in the alcohol-only use group (21.9%). Figure 2 suggests that dual use group reported a higher prevalence of IDR compared to no use and single use groups, with approximately half of adolescents who reported dual use reporting IDR. Additionally, around 10% of dual use group were unsure whether they had experienced IDR, which is higher than the proportion among no use and single use groups.

Figure 1. Percentage of students who reported exclusive alcohol impaired driving or riding with an alcohol-impaired driver (alcohol-IDR), exclusive cannabis-IDR (cannabis-IDR), both alcohol- and cannabis-impaired-IDR (alcohol-cannabis-IDR) and were unsure if they have experienced either alcohol or cannabis-IDR.

Figure 2. Percentage of students who reported exclusive alcohol and/or cannabis impaired driving or riding with an impaired driver (IDR) or were unsure if they have experienced either alcohol- and/or cannabis-IDR by reported substance use.

Table 1. Adjusted odds ratios for engaging in exclusive alcohol-IDR, exclusive cannabis-IDR, and alcohol-cannabis-IDR, 2021–22 COMPASS study.

	Drove under the influence or rode with a driver under influence of
	Exclusive alcohol-IDR* (n = 5,245)	Exclusive cannabis-IDR (n = 2,246)	Alcohol-cannabis-IDR** (n = 2,803)
Substance use (none as ref)
Alcohol only	2.72 (2.36, 3.14)‡	2.02 (1.57, 2.61)	3.91 (3.19, 4.78)
Cannabis only	1.40 (0.91, 2.15)	11.84 (8.73, 16.03)	11.53 (8.72, 15.26)
Dual use	2.70 (2.06, 3.53)	12.76 (9.96, 16.35)	20.85 (13.52, 29.78)
Gender
Cisgender girl
Alcohol only	2.87 (2.54, 3.26)	2.35 (1.86, 2.98)	4.18 (3.33, 5.25)
Cannabis only	1.36 (0.75, 2.46)	13.54 (9.20, 19.94)	12.87 (8.34, 19.86)
Dual use	2.99 (2.28, 3.90)	14.95 (11.41, 19.58)	11.94 (4.88, 21.00)
Cisgender boy
Alcohol only	2.90 (2.51, 3.35)	2.61 (1.96, 3.47)	4.55 (3.42, 6.04)
Cannabis only	0.98 (0.48, 2.01)	15.42 (10.08, 23.59)	18.01 (11.62, 27.92)
Dual use	3.38 (2.46, 4.64)	19.03 (13.81, 26.23)	34.09 (20.42, 52.39)
Gender diverse
Alcohol only	2.48 (1.64, 3.75)	1.16 (0.54, 2.51)	3.56 (2.13, 5.94)
Cannabis only	1.89 (0.71, 5.03)	9.17 (4.54, 18.50)	11.00 (5.72, 21.20)
Dual use	1.84 (0.85, 4.00)	6.65 (3.31, 13.36)	14.29 (1.35, 35.08)
Age
15 years old or younger
Alcohol only	2.76 (2.34, 3.25)	1.97 (1.37, 2.85)	4.84 (3.71, 6.31)
Cannabis only	1.38 (0.78, 2.43)	13.31 (8.94, 19.83)	17.07 (11.54, 25.24)
Dual use	2.84 (2.01, 4.03)	12.95 (9.02, 18.60)	20.94 (18.91, 46.70)
16 years old or older
Alcohol only	2.55 (1.92, 3.40)	2.14 (1.52, 3.02)	3.03 (2.22, 4.14)
Cannabis only	1.35 (0.73, 2.48)	9.35 (6.36, 13.75)	7.77 (5.20, 11.62)
Dual use	2.74 (1.86, 4.03)	12.42 (9.18, 16.81)	16.26 (10.08, 24.33)

* IDR: impaired driving or riding with an impaired driver.

**Alcohol-cannabis-IDR: both alcohol- and cannabis-IDR.

^‡ Significant odds ratios at p < 0.05 are bolded.

The prevalence of IDR among different demographic groups is shown in Supplementary Table A2. A higher proportion of girls reported IDR compared to boys. Gender-diverse students, on the other hand, reported a higher prevalence of alcohol-cannabis-IDR than both girls and boys. The occurrence of exclusive alcohol-IDR decreased with age, declining from 9.5% among 13-year-old students to 6.3% among those aged 17 and older. In contrast, the prevalence of exclusive cannabis-IDR and alcohol-cannabis-IDR increased with age, with reports among students aged 17 and older being more than three times higher than those among 13-year-olds.

Table 2. Adjusted odds ratios for engaging in exclusive alcohol-IDR, exclusive cannabis-IDR, and alcohol-cannabis-IDR among a subgroup of adolescents who reported monthly substance use, 2021–22 COMPASS study.

	Exclusive alcohol-IDR* (n = 5,245)	Exclusive cannabis-IDR (n = 2,246)	Alcohol-cannabis-IDR (n = 2,803)
Overall
Dual vs alcohol-only use (ref.)	1.03 (0.85, 1.25)	6.11 (4.99, 7.47)	9.51 (8.28, 10.92)
Dual vs cannabis-only use (ref.)	2.00 (1.41, 2.84)*	1.08 (0.86, 1.35)	2.99 (2.44, 3.66)
Gender
Cisgender girl
Dual vs exclusive alcohol use	1.08 (0.91, 1.30)	6.26 (5.25, 7.48)	7.30 (6.23, 8.55)
Dual vs exclusive cannabis use	2.61 (1.62, 4.20)	1.02 (0.79, 1.30)	2.21 (1.66, 2.96)
Cisgender boy
Dual vs exclusive alcohol use	1.16 (0.95, 1.43)	6.69 (5.45, 8.20)	10.79 (9.04, 12.88)
Dual vs exclusive cannabis use	3.15 (1.79, 5.56)	1.16 (0.89, 1.52)	3.31 (2.46, 4.46)
Gender diverse
Dual vs exclusive alcohol use	0.94 (0.54, 1.63)	6.01 (3.40, 10.63)	10.20 (6.98, 14.90)
Dual vs exclusive cannabis use	0.84 (0.37, 1.91)	1.04 (0.59, 1.85)	2.94 (1.78, 4.80)

* Significant odds ratios at p < 0.05 are bolded.

When examining the relationship between IDR and ethnicity, we found that a higher percentage of students who self-identified as White reported exclusive alcohol-IDR compared to other ethnic groups. Conversely, a higher proportion of Black students reported alcohol-cannabis-IDR. Among the ethnic groups surveyed, East Asian students reported the lowest percentage of the three IDRs.

Furthermore, there was a notable difference based on perceived family affluence. A higher percentage of students who perceived their family affluence as less comfortable than their peers reported both exclusive cannabis-IDR and alcohol-cannabis-IDR. In terms of geographic location, students in large cities reported lower rates of the three IDRs compared to those in small urban/rural areas. The results indicate that increased weekly expenditure was associated with higher ORs of cannabis-IDR and alcohol-cannabis-IDR reporting.

Association between IDR and dual use compared to no use

Table 1 presents the adjusted odds ratios for exclusive alcohol-IDR, exclusive cannabis-IDR, and alcohol-cannabis-IDR among adolescents who reported substance use compared to those reported no use. All covariates demonstrated significance in the univariate models and were subsequently incorporated into the multivariable analysis, indicating that the adjusted odds ratios presented in Table 1 are adjusted for all covariates. The findings indicate that alcohol-only use was associated with higher odds of reporting the three IDRs compared to no use. On the other hand, cannabis-only use was not associated with exclusive alcohol-IDR, but it increased the odds of exclusive cannabis-IDR and alcohol-cannabis-IDR compared to no use.

In the models, the interaction terms of substance use, and gender and age were significant therefore we ran stratified analyses. When examining the results of the stratified analysis by gender, we observed a weaker association between substance use and IDR for gender diverse adolescents compared to girls and boys. The results of age stratified analysis indicate that substance use in early adolescence (aged 15 and below) was slightly stronger association with reporting IDR.

Association between IDR and dual use compared to single use

Table 2 displays the adjusted odds ratios for the three IDRs comparing dual use to alcohol-only or cannabis-only use. The results reveal a significant increase in the odds of alcohol-cannabis-IDR among adolescents who reported dual use. Additionally, dual use was associated with higher odds of exclusive cannabis-IDR among alcohol-only use, indicating an increased likelihood of riding with a cannabis-impaired driver. Similarly, dual use significantly raised the odds of riding with an alcohol-impaired driver among the cannabis-only use group.

Considering the significant interaction between substance use and gender in the multivariable models, we conducted stratified analyses for gender groups as shown in Table 2. The results indicate that the association between dual use and alcohol-cannabis-IDR was stronger among adolescents who identified as gender diverse and boys than girls. On the other hand, the interaction terms between substance use and age (15 or under and 16 and above) were not found to be significant in the models, suggesting that the association between substance use and IDR did not differ significantly between these two age groups.

Discussion

The present research sought to assess whether dual use of alcohol and cannabis was associated with increased odds of IDR compared to use of alcohol or cannabis individually among adolescents. Our results indicate a higher prevalence of alcohol-related IDR compared to cannabis-related IDR, with the number of exclusive alcohol-IDR being more than twice of exclusive cannabis-IDR. This finding underscores that alcohol is the more commonly involved substance in IDR. However, it is important to note that the rate of experiencing cannabis-IDR was higher among those who reported cannabis-only use than the rate of alcohol-IDR among those who reported alcohol-only use. This indicates that adolescents who use cannabis are more likely to experience IDR than adolescents who only use alcohol.

Our study suggests that adolescents who reported using both alcohol and cannabis in the past month had significantly higher odds of IDR compared to those who solely used alcohol or cannabis. This finding is consistent with prior research (Terry-McElrath et al. 2014; Yurasek et al. 2017), which also indicated an increased likelihood of impaired driving associated with concurrent alcohol and cannabis use. The increased risk may be partially attributed to the synergistic effects observed when both substances are used simultaneously, as evidenced in previous research (Davis et al. 2019; Stevens et al. 2021; Simmons et al. 2022). However, it is important to note that our definition of dual use includes instances where alcohol and cannabis use may have occurred at different times within the past month. The increased odds of IDR associated with no-simultaneous use could be attributed to several factors. First, these individuals are more likely to be around individuals who use the other substance, potentially influencing their own engagement in risky transportation behaviors (Hultgren et al. 2021). Additionally, no-simultaneous use may increase the amount and frequency of alcohol and cannabis use (Yurasek et al. 2017), subsequently increasing the likelihood of experiencing impaired driving incidents. Thus, both the social influence and the escalation of substance use may contribute to an increased risk of IDR among adolescents who do not use substance simultaneously.

Consistent with the patterns of substance use among adolescents, where a greater number of students reported alcohol use compared to cannabis use, our findings indicate a higher prevalence of alcohol-IDR than cannabis-IDR. This suggests that alcohol continues to be the primary impairing substance for drivers, which is consistent with previous research studies (Hels et al. 2013; Martin et al. 2017). This trend can potentially be explained by several factors. Firstly, alcohol has been legally available for a longer period of time compared to cannabis, which was only legalized in 2018 in our study jurisdictions. Consequently, a larger number of individuals consume alcohol, increasing the risk of IDR. Additionally, alcohol is more socially accepted relative to cannabis, potentially leading to a higher occurrence of alcohol-IDR compared to cannabis-IDR. Because of this larger number of alcohol use, the risk of riding with an alcohol-impaired driver increases. Particularly, riding with a parent/guardian or sibling who is alcohol-impaired could be more likely than riding with an impaired driver under the influence of cannabis. Furthermore, detecting alcohol-impaired driving for passengers is typically easier than detecting cannabis-impaired driving, resulting in a higher reporting of alcohol-IDR. Additionally, the use of cannabis is often accompanied by the consumption of alcohol, thereby increasing the likelihood of alcohol-IDR (Martin et al. 2017).

In our sample, although there were more instances of alcohol-IDR, a higher proportion of adolescents who reported cannabis use engaged in cannabis-IDR compared to those who reported alcohol use. This findings consistent with previous research (Hultgren et al. 2021) and can be partially attributed to a perceived lower risk associated with cannabis-IDR in comparison to alcohol-IDR (Erin Goodman et al. 2020; Cantor et al. 2021; Donnan et al. 2022). The legalization and decriminalization of cannabis in Canada, along with a relatively short history of public awareness campaigns regarding the risks linked to cannabis-IDR, may contribute to this perception of reduced risk associated with cannabis use. This observation highlights the need for ongoing monitoring of trends in cannabis IDR, particularly as cannabis use becomes more socially acceptable and readily available in communities. It is possible that we may observe an increase in cannabis IDR and a corresponding decrease in alcohol IDR over time. Continued vigilance and research in this area are crucial to better understand and address the potential shifts in IDR patterns associated with the changing landscape of cannabis use.

One notable finding from our study is that a significant proportion of students expressed uncertainty regarding whether they had engaged in impaired driving or rode in a vehicle with an impaired driver. There are several potential explanations for this uncertainty. First, since a substantial number of participants were below the legal driving age, it is plausible that they were uncertain about the impairment status of the person they were riding with, such as their parents, older siblings, or friends (Leadbeater et al. 2008; Harris et al. 2017; Camenga et al. 2023). Additionally, the lack of knowledge or consensus regarding the appropriate waiting time after various forms of cannabis use could contribute to this uncertainty. Unlike alcohol, there is currently conflicting recommendations on the duration of impairment after cannabis use and methods of detecting cannabis impaired driving (Ginsburg 2019; Hasan et al. 2022); therefore adolescents may be unclear whether they are driving under the influence of cannabis or riding with a cannabis-impaired driver. Regardless of the underlying reasons for this uncertainty, our findings emphasize the importance of addressing this issue as a significant proportion of adolescents may be unknowingly putting their lives at risk. To address this issue, it is imperative to implement educational programs targeting adolescents to enhance awareness regarding impairment resulting from substance use, encompassing both alcohol and cannabis. These campaigns should prioritize educating adolescents on recognizing signs of impairment in themselves and others, along with highlighting the risks associated with IDR. Additionally, the development of standardized methods for assessing impairment due to substance use, particularly cannabis, is crucial. These measures will empower adolescents to make informed decisions regarding driving or riding with impaired individuals.

As evident from the literature (Harris et al. 2017; Camenga et al. 2023), a substantial proportion of students who report riding with an impaired driver indicate that the driver was an adult in the home. This could be the case in our sample, although we lack data to confirm this in our study. The impact is significant, as parental impaired driving poses an immediate risk of a crash. Further, children’s normative perceptions of driving or riding impaired, as observed in their parents, have a lasting impact on their IDR behavior (Leadbeater et al. 2008; Camenga et al. 2023). This emphasizes the significance of involving parents in preventive programs, screening, and counseling regarding IDR (Harris et al. 2017). It also highlights the need to implement family-centered approaches to mitigate the risk of teen driving. In this study, however, there is a limitation in our ability to ascertain whether the impaired driver whom adolescents ride with is a parent.

Our study revealed an intriguing pattern regarding the reporting of IDR among different gender groups. Contrary to our hypothesis and past research that suggested higher prevalence of IDR among male adolescents (Li et al. 2013; Cantor et al. 2021), our findings indicate that girls in our sample were more likely to report exclusive alcohol, exclusive cannabis, and both alcohol-and cannabis-IDR than boys. This contrary finding can be attributed to the inclusion of both impaired driving and riding with an impaired driver in our measure. Research has shown that female students are more prone to riding with impaired drivers, including parents (Harris et al. 2017), than males (Minaker et al. 2017). Therefore, the higher proportion of IDR reported by girls in our study may be a result of their increased exposure to riding with impaired drivers. This explanation becomes more apparent considering that many of our participants are underage and do not have their driver’s licenses, and thus, their reports may primarily be related to riding with an impaired driver. Concerningly, students who identified as gender diverse were more likely to report both alcohol- and cannabis-IDR than cisgender girls and boys. However, it is important to note that the sample size of the gender diverse group was relatively small, leading to wider confidence intervals in the estimates of odds ratios. Therefore, future research that deliberately oversamples diverse gender groups is necessary to investigate potential variations among more refined gender categories. By expanding our understanding of the nuanced relationship between gender and IDR, we can develop targeted interventions and prevention strategies to address the specific needs and risks associated with gendered roles and norms.

Younger students were more likely to report alcohol-IDR, despite not being eligible to drive themselves. As explained in the previous paragraph and consistent with past research (Cartwright and Asbridge 2011), this finding suggests that younger students may be more likely to ride with individuals who are impaired than to drive impaired themselves. This highlights the importance of considering not only the behaviors of young drivers but also the behaviors of those who provide transportation to underage individuals. Effective prevention efforts should not only focus on educating young drivers about the dangers of impaired driving but also emphasize the importance of having designated drivers who abstain from substance use to ensure the safety of passengers. Another plausible explanation for age effect could be a cohort effect, wherein the younger demographic in our study is growing up during a period when cannabis is legally accessible, compared to older adolescents who grown up either prior to its legalization or during the early stages of legalization (Donnan et al. 2022).

The findings of our study support results of past research that suggests IDR is more prevalent in rural and small urban areas compared to large cities (Carpino et al. 2020). This difference can be attributed to several factors, including the availability of public transit and active travel (e.g., cycling infrastructure) options and more walkable areas (e.g., closer destinations of interest and mixed land use). While individuals in rural areas heavily depend on their cars for transportation, large cities typically offer more extensive public transportation systems, which may help reduce the necessity of driving or riding under the influence. Furthermore, a qualitative research study conducted among Canadian youth has revealed that the relatively lower allocation of law enforcement resources in rural and small urban areas may contribute to an increase IDR in these areas (Donnan et al. 2022). The authors have also highlighted a prevailing belief among participants, particularly those residing in rural areas, that driving shorter distances is perceived as being less risky compared to longer distances.

While our study provides valuable insights into IDR among adolescents, it is important to consider several limitations when interpreting the results. Firstly, IDR was measured through self-report, which introduces the potential for recall bias and individuals’ subjective perception of their impairment or the driver’s impairment. Moreover, the use of self-report data is subject to social desirability bias. This bias can lead participants to underreport or withhold information about their involvement in IDR, particularly as an illegal behavior. Secondly, our study did not differentiate between driving under the influence and riding with an impaired driver, preventing us from examining the specific contribution of impaired drivers versus passengers with impaired drivers. These two behaviors signify distinct actions. In the latter scenario, the identity of the driver remains unknown, leaving room for speculation that it could have been a peer, parent/guardian, other family member (such as an older sibling), or another individual, each with potentially distinct implications. Younger adolescents riding with an impaired parent or guardian involves an uncontrolled behavior that significantly differs from impaired driving. Future research should explore factors associated with impaired driving versus riding with impaired drivers and assess the relationship of the driver to the adolescent (i.e., parent/guardian, other family member, peer, etc.), to gain a more comprehensive understanding of the issue. Relatedly, there is a lack of data in this study regarding whether participants were licensed drivers or had access to a vehicle, thus limiting our ability to comprehend their influence on impaired driving or riding. Thirdly, our study did not investigate the proportion of IDR instances that occurred under the simultaneous use of alcohol and cannabis. Further studies should aim to differentiate between simultaneous alcohol and cannabis IDR and separate episodes to elucidate potential differences in risk factors and outcomes. Likewise, while our research examined the two most commonly used substances, co-use and IDR involving other substances, such as opioids or methamphetamines, warrant consideration. Another limitation is that the timeframe for substance use (at least once a month in the last year) and IDR experience (in the past month) were different, which may influence the evaluation of monthly IDR behavior and substance use. Lastly, missing data may introduce bias, as students engaging in substance use and/or IDR may have been more likely to not respond to these items as in our data 5521 (7.3%) did not respond to alcohol- or cannabis-IDR and were excluded from the analysis.

The findings of our study highlight the elevated risk of IDR among adolescents that use cannabis and alcohol, particularly those who use both substances. Targeted interventions about the risks associated with both cannabis- and alcohol-IDR should be explored, although even some adolescents who had not used either substance appear to be putting themselves at risk by riding in a car with someone under the influence. Further, a concerning proportion of adolescents are unsure of whether they have experienced IDR, suggesting interventions to improve understanding of how to assess impairment may be warranted. Continued research is needed to explore gendered influences and perceptions of substance-specific IDR risk, giving potential shifts in social norms.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The COMPASS study has been supported by a bridge grant from the CIHR Institute of Nutrition, Metabolism and Diabetes (INMD) through the “Obesity – Interventions to Prevent or Treat” priority funding awards (OOP-110788; awarded to STL), an operating grant from the CIHR Institute of Population and Public Health (IPPH) (MOP-114875; awarded to STL), a CIHR project grant (PJT-148562; awarded to STL), a CIHR bridge grant (PJT-149092; awarded to KAP/STL), a CIHR project grant (PJT-159693; awarded to KAP), and by a research funding arrangement with Health Canada (#1617-HQ-000012; contract awarded to STL), a CIHR-Canadian Center on Substance Abuse (CCSA) team grant (OF7 B1-PCPEGT 410-10-9633; awarded to STL). The COMPASS-Quebec project additionally benefits from funding from the Ministère de la Santé et des Services sociaux of the province of Québec, and the Direction régionale de santé publique du CIUSSS de la Capitale-Nationale. KAP is the Canada Research Chair in Child Health Equity and Inclusion.