Does Believing Alcohol Causes Cancer Moderate the Relationship Between Consumer Awareness of the Alcohol–Cancer Link and Support for Alcohol Policies? Findings From a Canadian Cross‐Sectional Study

ABSTRACT

Introduction

Extending research observing an association between awareness that alcohol causes cancer and support for alcohol policies, this study examined if believing or accepting alcohol causes cancer moderates the relationship between awareness of alcohol as a carcinogen and policy support.

Methods

Adult alcohol consumers (n = 5180) in Canada completed an online survey in March–April 2023. Four separate logistic regression models were conducted with policy support affecting alcohol availability, pricing, marketing and labelling as outcomes to assess if believing alcohol causes seven types of cancer moderates the relationship between awareness of the alcohol–cancer link and support for alcohol policies. An interaction between awareness and belief was included as a predictor, adjusting for covariates.

Results

Overall, 29.3% were aware alcohol causes seven types of cancer and, of those aware, 83.6% believed this link. Those both aware of and believing that alcohol causes cancer had higher odds of supporting policies restricting alcohol availability (OR 1.76, 95% CI 1.13, 2.74) and marketing (OR 1.75, 95% CI 1.16, 2.64) than those not aware and did not believe. Consumers who were both aware of and believed the alcohol–cancer link had higher odds of supporting labelling policies (OR 1.59, 95% CI 1.05, 2.40), although this was not significant after adjusting for multiple comparisons.

Discussion and Conclusions

This study highlights that believing alcohol is a carcinogen moderates the relationship between awareness of the alcohol–cancer link and support for policies restricting alcohol availability and marketing. Future longitudinal studies are needed to test interventions for effectively raising awareness and strengthening belief and acceptance of alcohol‐related cancer risks.

Summary.

• Believing alcohol causes cancer moderates the relationship between consumer awareness of alcohol as a carcinogen and support for alcohol availability and marketing policies.

• Results also indicate believing alcohol causes cancer moderates the relationship between consumer awareness and support for alcohol labelling policies, although this did not remain significant after accounting for multiple comparisons.

• These findings highlight consumer awareness of alcohol as a carcinogen may not be sufficient and further interventions to enhance belief or acceptance that alcohol is carcinogenic may be required to strengthen public support for alcohol policies.

• Future studies should conduct longitudinal, real‐world research to investigate if public health information‐based interventions communicating alcohol–cancer messaging can effectively raise consumer awareness, increase knowledge and gain acceptance over time.

1. Introduction

Alcohol is consumed each year by two billion people worldwide, with global consumption expected to increase 17% by 2030 [1]. In 2019, 38% of consumers worldwide reported heavy episodic drinking or ‘binge drinking’, defined as drinking at least 60 g of pure alcohol on one or more occasions in the past month [2]. Alcohol is a leading risk factor for disease and premature death [3] and is a component cause of 230 acute and chronic health conditions, including seven cancer types [3, 4]. There is a dose–response relationship between alcohol and increased cancer risk, with even light to moderate consumption increasing cancer risk [5] leading the World Health Organization (WHO) to caution that no level of alcohol use is safe because of this risk [6]. In 2020, alcohol accounted for an estimated 741,300 new cancer cases globally [7], including 7000 new cases in Canada. In addition, the economic impacts of alcohol, including healthcare costs and lost productivity, are substantial, exceeding 1% of the gross national product in high‐ and middle‐income countries [8].

Public awareness that alcohol causes cancer is low worldwide [9], with only 13% in some jurisdictions aware of the link between alcohol and cancer [9]. Findings from surveys in Canada, the United States and Australia suggest public awareness that alcohol is carcinogenic ranged between 26% and 56%, depending on the cancer type [10, 11, 12]. The WHO argues one reason for low public awareness of alcohol‐related health risks, particularly cancer, is the alcohol information environment is saturated with industry messages that alcohol is acceptable and safe [13]. On the other hand, relatively little public health information‐based interventions about alcohol have been implemented. This is problematic as increased public awareness of the link between alcohol and cancer has been associated with stronger public support for policies affecting alcohol price, marketing and availability [14, 15, 16, 17, 18, 19]. Coordinated, comprehensive and effective policies are needed to reduce the health and economic burden of alcohol. International evidence consistently indicates alcohol policies have a protective effect against per capita alcohol consumption and harms [20, 21, 22], with stricter alcohol policy environments capable of shifting a population towards lower alcohol harms [20]. A large robust evidence base suggests restricting the availability and marketing and regulating the price of alcohol are the most cost‐effective measures for reducing population‐level alcohol consumption and harms [13, 23], yet many of these policies are not well‐supported by the public [14, 15, 16, 17, 18] and hence policymakers, and are poorly implemented.

Fitzgerald et al. [24] argue a key first step to achieving a sustainable reduction in alcohol use and attributable harm, is to change the social environment (e.g., acceptability, norms), and public perceptions of alcohol health risks. Information‐based interventions that increase public awareness of alcohol health risks, particularly the causal link between alcohol and cancer, are promising approaches for balancing the information environment with messaging about the potential impact of alcohol on health and safety and creating a more supportive social climate for increasing public approval of effective alcohol control policies [2, 14, 15, 16, 17]. The Elaboration Likelihood Model indicates for a health message to have an impact, a message must generate exposure, attract attention, and be processed by the target audience [25]. Frequent exposure to a message increases the probability that an audience will learn and internalise the message as each exposure provides an additional opportunity for message processing [26]. Repeated exposure can also lead to priming effects, making individuals more open to considering attitude and behavioural changes. This receptiveness increases as the message becomes more familiar, decreasing resistance and enhancing openness to the message [27]. In addition, repeated exposure to a message can lead to discussions about the message, facilitating the message's reach beyond the initial audience [28, 29]. If a message is successful in those early‐stage processes, later‐stage processes such as attitude and perception changes may occur. Accordingly, affecting these early‐stage processes is from a theoretical perspective, a prerequisite for attitude and perception changes. In the absence of such processes, it is unlikely that health messages will have subsequent population‐level effects. Extending existing research observing an association between public awareness that alcohol causes cancer and support for alcohol policies [14, 15, 16, 17, 18, 19], the primary aim of this study was to examine if believing or accepting the alcohol–cancer link moderates the relationship between awareness that alcohol causes cancer and support for alcohol policies among a large sample of alcohol consumers in Canada (see Figure S1, for conceptual diagram).

2. Methods

2.1. Study Design and Participants

Participants were part of a larger online study conducted in March–April 2023 designed to experimentally test the influence of alcohol container labels that has been previously described [30]. Participants were residents in one of 10 Canadian provinces, between the legal drinking age (18 or 19 years) and 64 years, were a current alcohol consumer (≥ 1 drink in past 30 days), did not report being pregnant or breastfeeding, understood English or French, and had access to the internet.

Participants were recruited through the Leger Marketing online participant panel, a survey‐sampling company that consists of approximately 400,000 individuals in Canada [31]. The Leger panel participants were recruited through probability and non‐probability sampling methods. For this study, Leger used quota‐based sampling with the distribution proportional to each of the 10 provincial populations by sex and age group, based on the 2021 Canadian Census [32]. Leger emailed panellists an invitation to participate in the study and with access to the online survey via a hyperlink. Participant eligibility was screened according to the inclusion criteria about alcohol consumption in the prior 30 days and pregnancy or breastfeeding. Participants who completed the survey were compensated according to the panel's remuneration structure. Surveys were conducted in English or French. The full study was estimated to take approximately 20 min to complete. This study was approved by the Research Ethics Board at Public Health Ontario (ID 2022‐0.30.02) and all participants provided electronic informed consent.

In total, 17,260 consented and were screened for eligibility and 5500 were eligible and randomised. Participants were excluded if they were duplicates (n = 7), completed the survey in < 25% of the median time (n = 85), withdrew consent (n = 7), failed data checks (could not correctly identify the current month from a multiple choice question; n = 140), responded ‘No’ when asked if they were honest about their alcohol use (n = 37) or had a pattern of ‘Don't know/Prefer not to say’ responses (n = 44). In total, 5180 participants were included in the final sample.

2.2. Measures

2.2.1. Awareness and Belief That Alcohol Causes Cancer

To assess awareness and belief that alcohol causes cancer, participants were presented with the statement, ‘Alcohol causes at least seven different types of cancer’ and subsequently asked ‘Is this information new to you?’ and ‘Do you believe this information?’ Response options included ‘No’, ‘Yes’, ‘Don't know’ and ‘Prefer not to say’. To make it easier to interpret, the question ‘Is this information new to you?’ was rephrased to, ‘Are you aware of this information?’ in the present analysis. Therefore, participants who answered ‘No’ and ‘Yes’ to this information being new to them were categorised into the ‘Yes’ and ‘No’ categories that they were aware of this information, respectively.

2.2.2. Support for Alcohol Policies

To assess support for alcohol policies, participants were asked, ‘To reduce the problems associated with drinking alcohol, to what extent do you support or oppose:’, followed by a list of 12 policy items within four policy domains affecting alcohol availability, pricing, marketing and labelling (Figure 1). Policies were presented in random order, and participants were asked to indicate the extent of their support or opposition on a five‐point scale (1 = Strongly oppose, 5 = Strongly support), with ‘Don't know’ and ‘Prefer not say’ included as options. This measure was adapted from the Australian National Drug Strategy Household survey and has been used in previous studies assessing associations between awareness of the alcohol–cancer link and policy support [16, 18].

FIGURE 1.

Level of support for alcohol policies (N = 5180). Percent missing varies from 0.04% to 0.15% across the 12 policy measures.

2.2.3. Socio‐Demographics

Socio‐demographics included age, gender, province of residence, education, sexual orientation and race/ethnicity (see Table 1 for coding options).

TABLE 1.

Sample characteristics (N = 5180).

Characteristic	N (%)
Age group
Younger adults (18/19–40 years)	2623 (50.6)
Older adults (41–64 years)	2557 (49.4)
Gender
Women	2637 (51.0)
Men	2468 (47.7)
Gender‐fluid/non‐binary/transgender/two‐spirit/other	65 (1.3)
Missing	10
Province of residence
Ontario	1992 (38.5)
Quebec	1185 (22.9)
Prairie provinces (Alberta, Saskatchewan, Manitoba)	955 (18.4)
British Columbia	710 (13.7)
Eastern/Maritime provinces	338 (6.5)
Awareness of Canada's Guidance on Alcohol and Health
No/Don't know	2826 (54.6)
Yes	2351 (45.4)
Missing	3
Education
High school or below/trades/college/some university	2746 (53.4)
Bachelor degree or above	2401 (46.6)
Don't know/Missing	33
Sexual orientation
Heterosexual	4147 (83.1)
Asexual/bisexual/gay/lesbian/pansexual/queer/two‐spirit/other	844 (16.9)
Don't know/Missing	189
Race/ethnicity
White only	3754 (74.5)
Asian (East/southeast Asian, South Asian)	727 (14.4)
Mixed/Other (including Black, Latino, Middle Eastern, Indigenous)	559 (11.1)
Don't know/Missing	140
Preferred alcohol type
Wine	1584 (30.9)
Beer (regular or light beer)	1601 (31.2)
Spirits (e.g., rum, whisky, vodka, mixed drinks)	1179 (23.0)
Cider/coolers/hard seltzers/other	770 (15.0)
Don't know/Missing	46
Hazardous alcohol use (AUDIT‐C) a
No (< 3/4 AUDIT‐C score)	2079 (40.7)
Yes ($\ge$ 3/4 AUDIT‐C score)	3024 (59.3)
Missing	77

Abbreviation: AUDIT‐C, Alcohol Use Disorders Identification Test‐Concise.

^a AUDIT‐C score: ≥ 3 (women or other)/≥ 4 (men) identifies hazardous use or active alcohol use disorder.

2.2.4. Other Covariates

Other covariates measured in this study included awareness of national alcohol guidance in Canada, preferred alcohol type, and hazardous alcohol use as measured with the three‐item Alcohol Use Disorders Identification Test‐Concise (AUDIT‐C). Participants were provided with a photo of a standard drink guide to aid accurate responding.

Awareness of national alcohol guidance was measured by asking, ‘Before today, had you heard of Canada's Guidance on Alcohol and Health?’ Response options included ‘No’, ‘Yes’, ‘Don't know’ and ‘Prefer not to say’. Preferred alcohol type was measured by asking, ‘When you drink alcohol, what type do you usually drink?’, categorised as wine, beer (regular or light beer), spirits (e.g., rum, whisky, vodka, mixed drinks), cider/coolers/hard seltzers/other, and Don't know/Prefer not to say. Hazardous alcohol use was assessed by calculating the score from three questions abstracted from the AUDIT‐C. Participants were asked, during the past 12 months: ‘How often did you have a drink containing alcohol?’ (Response options: ‘Less than once a month’, ‘Once a month’, ‘2 to 3 times a month’, ‘Once a week’, ‘2 to 3 times a week’, ‘4 to 5 times a week’, ‘Daily or almost daily’, ‘Don't know’, ‘Prefer not to say’), ‘On those days when you have drank alcohol, how many drinks did you usually have?’ (Response options: ‘1 or 2’, ‘3 or 4’, ‘5 or 6’, ‘7’, ‘8’, or ‘9, 10 or more’, ‘Don't know’, ‘Prefer not to say’) and ‘How often have you had 4 (female)/5 (male) or more drinks on one occasion?’ (Response options: Never in the past 12 months, Less than monthly, Monthly, Weekly, Daily or almost daily, Don't know, Prefer not to say). An AUDIT‐C score of ≥ 3/4 for women/men, respectively identifies hazardous use.

2.3. Statistical Analysis

The analyses involved six steps. First, to determine if policy measures within each of the four policy domains could be combined, the reliability of internal consistency was assessed using the Cronbach alpha coefficient (α). Values for α range from 0 to 1, with higher values indicating better reliability. Good reliability was established for all four alcohol policy domains (Cronbach α = 0.806 for availability, 0.708 for pricing, 0.862 for marketing and 0.874 for labelling) and thus, the measures in each domain were combined into a single outcome for availability, pricing, marketing and labelling. Second, responses for support within each policy domain were dichotomised into ‘Support’ (either Strongly support or Support any of the policy measures within each domain) or ‘Do Not Support’ (either Strongly oppose, Oppose, Neutral or Don't know to all of the policy measures within each domain) as done in previous research [18]. Third, logistic regression examined associations between sociodemographic and alcohol use behaviours and awareness that alcohol causes cancer, with the awareness dichotomised as ‘Yes’ and ‘No/Don't know’ [18]. Fourth, multinomial logistic regression examined associations between covariates, mentioned in step three, and believing alcohol causes cancer, categorised as ‘Yes’, ‘Don't know’ and ‘No’. Fifth, to assess the association between awareness that alcohol causes cancer and support for alcohol polices, four separate logistic regression models were conducted with the four policy domains, availability, pricing, marketing and labelling, as outcomes and awareness as a predictor, adjusting for covariates mentioned in step three. Lastly, to investigate if believing alcohol causes cancer moderates the relationship between awareness of the alcohol–cancer link and support for alcohol policies, four separate logistic regression models were conducted with the four policy domains mentioned in step five as outcomes. An interaction between awareness and belief that alcohol causes cancer was included as a predictor to test for moderation, adjusting for covariates mentioned in step three. Adjusted odds ratios (OR) and corresponding 95% confidence intervals (CI) were estimated to quantify associations in logistic regression models.

A sensitivity analysis evaluated the effect of excluding those that participated in the campaign ‘Dry February’ as the study was conducted in March and included measures assessing past 30‐day alcohol use, yet this did not substantially alter the main results (Supporting Information Tables). Statistical tests were two‐sided and evaluated at a 5% significance level. To reduce the risk of type 1 error, Benjamini–Hochberg adjustment of p‐values was used for multiple comparisons. Participants with ‘Don't know’ and ‘Missing’ response options for all covariates, and genders other than man/woman were excluded due to small sample sizes. The only exception was for awareness of national alcohol guidance, where ‘Don't know’ was grouped with ‘No’. ‘Prefer not to say’ responses were excluded from the outcome measures in all models and all outcome measures with > 10% in a category were analysed separately. Analyses were conducted using SAS Enterprise Guide 8.2 Update 1 (8.2.1.1223).

3. Results

3.1. Sample Characteristics

Sample characteristics are presented in Table 1. More than half of participants were younger adults aged 18/19–40 (50.6%) and female (51.0%). Approximately 4 in 10 participants were from Ontario (38.5%), Canada's most populous province. More than half of participants (54.6%) were either not aware or not sure of national alcohol guidance. More than half of participants had an education level lower than a Bachelor degree (53.4%). Majority of participants were heterosexual (83.1%) and approximately three‐quarters identified as White (74.5%). Beer was the most preferred alcohol type (31.2%). Over half of the sample (59.3%) had an AUDIT‐C score ≥ 3 or 4.

3.2. Awareness of Alcohol as a Carcinogen

Consumer awareness of the alcohol–cancer link in this sample was 29.3% (Figure 2a). Of those who were aware of this link, 83.6% believed this information, 11.7% did not believe and 4.5% were unsure (Figure 2b). Among those who were not aware or unsure of this information, 58.2% believed this information, 16.7% did not and 25.1% were unsure.

FIGURE 2.

(a) Awareness and belief of alcohol causing seven types of cancer (N = 5180). Number of participant exclusions for awareness (N = 11) and belief (N = 10) due to missing responses. (b) Belief that alcohol causes seven types of cancer among those aware (N = 1513) and unaware/Don't know (N = 3656) that alcohol causes cancer. Number of participant exclusions among those aware (N = 1) and unaware/Don't know (N = 5) alcohol causes cancer due to missing responses.

3.3. Associations Between Awareness That Alcohol Causes Cancer and Socio‐Demographic and Alcohol‐Related Characteristics

Higher odds of awareness of the alcohol–cancer link were observed among consumers who were aware of national alcohol guidance (OR 2.20, 95% CI 1.93, 2.52), with a higher level of education (OR 1.26, 95% CI 1.10, 1.44), and resided in British Columbia (OR 1.35, 95% CI 1.10, 1.64; Table 2) compared to Ontario. Participants identifying as Asian had lower odds of awareness of the alcohol–cancer link (OR 0.70, 95% CI 0.57, 0.86) compared to participants identifying as White. No other covariates were associated with awareness.

TABLE 2.

Associations between awareness and belief that alcohol causes cancer and key sociodemographic and alcohol use characteristics.

Exposure	Alcohol causes seven types of cancer
	Awareness (ref = No/Don't know) N = 4718	Belief (ref = No) N = 4717
	Yes	Yes	Don't know
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Age group
Older adults (41–64 years)	Ref	Ref	Ref
Younger adults (18/19–40 years)	1.10 (0.96, 1.26)	1.46 (1.22, 1.74)	0.75 (0.60, 0.93)
Gender
Men	Ref	Ref	Ref
Women	0.86 (0.75, 0.99)	1.75 (1.45, 2.10)	2.10 (1.69, 2.62)
Province of residence
Ontario	Ref	Ref	Ref
British Columbia	1.35 (1.10, 1.64)	1.32 (1.00, 1.73)	1.36 (0.98, 1.88)
Prairie provinces (Alberta, Saskatchewan, Manitoba)	1.09 (0.91, 1.30)	0.99 (0.79, 1.25)	1.08 (0.82, 1.42)
Quebec	1.08 (0.90, 1.30)	1.41 (1.11, 1.79)	1.12 (0.84, 1.48)
Eastern/maritime provinces	1.03 (0.78, 1.35)	1.07 (0.76, 1.51)	1.09 (0.72, 1.65)
Awareness of Canada's Guidance on Alcohol and Health
No/don't know	Ref	Ref	Ref
Yes	2.20 (1.93, 2.52)	1.45 (1.22, 1.72)	0.96 (0.77, 1.18)
Education
High school or below/trades/college/some university	Ref	Ref	Ref
Bachelor degree or above	1.26 (1.10, 1.44)	1.28 (1.07, 1.52)	0.90 (0.73, 1.12)
Sexual orientation
Heterosexual	Ref	Ref	Ref
Asexual/bisexual/gay/lesbian/pansexual/queer/two‐spirit/other	1.20 (1.01, 1.44)	1.01 (0.80, 1.27)	0.93 (0.70, 1.24)
Race/ethnicity
White only	Ref	Ref	Ref
Asian	0.70 (0.57, 0.86)	0.82 (0.64, 1.06)	0.78 (0.57, 1.07)
Mixed/other	0.84 (0.68, 1.04)	0.96 (0.73, 1.26)	0.82 (0.58, 1.15)
Preferred alcohol type
Wine	Ref	Ref	Ref
Beer (regular or light beer)	0.85 (0.72, 1.02)	0.82 (0.66, 1.04)	0.82 (0.62, 1.08)
Cider/coolers/hard seltzers/other	0.80 (0.65, 0.99)	0.89 (0.67, 1.19)	0.94 (0.67, 1.31)
Spirits (e.g., rum, whisky, vodka, mixed drinks)	0.88 (0.73, 1.05)	0.73 (0.58, 0.93)	0.76 (0.58, 1.02)
Hazardous alcohol use (AUDIT‐C) a
$\ge$ 3/4 AUDIT‐C score	Ref	Ref	Ref
< 3/4 AUDIT‐C score	0.96 (0.84, 1.10)	1.72 (1.44, 2.05)	1.54 (1.25, 1.91)

Note: Each model adjusted for all sociodemographics, awareness of national alcohol guidance and alcohol use behaviours.

Estimates in bold remained statistically significant at p < 0.05 after the Benjamini–Hochberg adjustment for multiple comparisons.

Abbreviations: aOR, adjusted odds ratio; AUDIT‐C, Alcohol Use Disorders Identification Test‐Concise; CI, confidence interval.

^a AUDIT‐C score: ≥ 3 (women)/≥ 4 (men) identifies hazardous use or active alcohol use disorder.

3.4. Associations Between Believing That Alcohol Causes Cancer and Socio‐Demographic and Alcohol‐Related Characteristics

Higher odds of believing alcohol causes cancer were found among younger adults (OR 1.46, 95% CI 1.22, 1.74) compared to older adults, women (OR 1.75, 95% CI 1.45, 2.10) compared to men, those with an AUDIT‐C score < 3/4 (OR 1.72, 95% CI 1.44, 2.05) compared to those with a score ≥ 3/4, those who were aware (OR 1.45, 95% CI 1.22, 1.72) compared to not aware of national alcohol guidance and those having an education at a Bachelor degree or above (OR 1.28, 95% CI 1.07, 1.52) compared to lower education levels (Table 2). Women (OR 2.10, 95% CI 1.69, 2.62) and those with an AUDIT‐C score < 3/4 (OR 1.54, 95% CI 1.25, 1.91) also had higher odds of reporting ‘Don't know’ for this link compared to men and those with an AUDIT‐C score ≥ 3/4, respectively. Lower odds of reporting ‘Don't know’ were observed among younger adults (OR 0.75, 95% CI 0.60, 0.93) compared to older adults.

3.5. Believing Moderates Awareness That Alcohol Causes Cancer and Support for Alcohol Policies

Those aware that alcohol causes cancer had higher odds of support for availability (OR 1.10, 95% CI 0.96, 1.25), pricing (OR 1.12, 95% CI 0.98, 1.29), marketing (OR 1.07, 95% CI 0.94, 1.22) and labelling (OR 1.08, 95% CI 0.93, 1.24) policies, although these results were not statistically significant (Table 3A). As shown in Table 3B, when belief is considered, those both aware of and believing that alcohol causes cancer had higher odds of support for availability (OR 1.76, 95% CI 1.13, 2.74) and marketing policies (OR 1.75, 95% CI 1.16, 2.64) compared to those who were not aware of the link and did not believe (see Figure S2, for a visual). Consumers who were both aware of and believed in the alcohol–cancer link had higher odds of support for labelling policies (OR 1.59, 95% CI 1.05, 2.40) compared to those who were not aware of the link and did not believe, although results were not significant following the Benjamini–Hochberg adjustment.

TABLE 3A.

Associations between awareness of the link between alcohol and cancer and support for alcohol policies.

Exposure	Support for alcohol policies
	Availability (N = 4715)	Pricing (N = 4714)	Marketing (N = 4715)	Labelling (N = 4705)
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Awareness
No/Don't know	Ref	Ref	Ref	Ref
Yes	1.10 (0.96, 1.25)	1.12 (0.98, 1.29)	1.07 (0.94, 1.22)	1.08 (0.93, 1.24)

Note: Each alcohol policy outcome modelled separately (four models), models adjusted for sociodemographic, awareness of national alcohol guidance, alcohol use behaviours and awareness/belief that alcohol causes seven types of cancer.

Abbreviations: aOR, adjusted odds ratio; CI, confidence interval.

TABLE 3B.

Associations with an interaction between awareness and belief of the alcohol and cancer link and support for alcohol policies.

Exposure	Support for alcohol policies
	Availability (N = 4709)	Pricing (N = 4708)	Marketing (N = 4709)	Labelling (N = 4699)
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Awareness × belief
Awareness (No/Don't know) × belief (No)	Ref	Ref	Ref	Ref
Awareness (Yes) × belief (Don't know)	1.54 (0.77, 3.09)	0.53 (0.21, 1.31)	1.26 (0.65, 2.44)	1.51 (0.79, 2.91)
Awareness (Yes) × belief (Yes)	1.76 (1.13, 2.74)	1.05 (0.67, 1.63)	1.75 (1.16, 2.64)	1.59 (1.05, 2.40)

Note: Each alcohol policy outcome modelled separately (four models), models adjusted for sociodemographic, awareness of national alcohol guidance, alcohol use behaviours and awareness/belief that alcohol causes seven types of cancer.

Estimates in bold remained statistically significant at p < 0.05 after the Benjamini–Hochberg adjustment for multiple comparisons.

Abbreviations: aOR, adjusted odds ratio; CI, confidence interval.

4. Discussion

Findings from this study suggest that believing that alcohol causes cancer moderates the relationship between consumer awareness that alcohol causes cancer and support for alcohol availability and marketing policies. Findings also show that believing alcohol causes cancer moderates the relationship between consumer awareness and support for alcohol labelling policies, although this relationship was not significant after adjusting for multiple comparisons. These findings highlight that awareness that alcohol causes cancer may not be sufficient for maximising public support for alcohol policies and further interventions that strengthen believing or accepting the alcohol–cancer link are warranted.

Approximately 30% of the study sample were aware that alcohol is a carcinogen, which is comparable to the low levels of awareness previously found in Canada [10]. We observed associations between higher educational attainment and awareness that alcohol causes cancer, which differ from previous research that identified additional sociodemographic characteristics associated with awareness, including being female, older age, higher socioeconomic status and lower alcohol use [16, 33]. This discrepancy in correlates may in part be due to the measure used in the current study assessing consumer awareness that ‘alcohol causes seven different types of cancer’ versus cancer more generally. Further, belief closely followed awareness, with the majority of participants believing that alcohol causes cancer, even among participants who reported being unaware of the alcohol–cancer link. This is a novel finding suggesting the potential for interventions designed to increase awareness to also directly contribute to strengthening consumer beliefs or acceptance that alcohol is carcinogenic. However, more than 40% of consumers who were unaware of the alcohol–cancer link reported not believing or being unsure about this relationship. Reluctance to believe may not be surprising given this information about alcohol and cancer risk is new and likely contradicts established beliefs among many alcohol consumers that alcohol is relatively safe. New, conflicting information can trigger psychological defence mechanisms that hinder the acceptance of new information [34]. This can be particularly pronounced when the new information challenges a person's identity and social behaviours. To reduce resistance to change beliefs and foster openness to alcohol and health risk messaging, interventions that incorporate frequent and repeated exposure to messages should be prioritised.

To improve public awareness that alcohol causes cancer, the WHO recommends government‐mandated health warning labels on product containers as one strategy for communicating alcohol‐related health risks to consumers at key points of intervention, including when purchasing, pouring and drinking alcohol [35]. Repeated exposure to health warning labels on alcohol containers is thought to affect consumer behaviour by gaining attention, eliciting aversive reactions and keeping the message in consumers' minds [36]. Evidence from real‐world settings indicates that repeated exposure to well‐designed health warning labels can raise consumer awareness of alcohol‐related cancer risks and reduce drinking intentions and alcohol use [37, 38, 39, 40]. Future studies are required to extend our understanding of the influence of health warning labels on consumer beliefs.

4.1. Strengths and Limitations

A strength of this study is the large sample size from Canadian provinces, which allowed adjustment for multiple risk factors, including socio‐demographics and alcohol behaviours. However, there were limitations. First, although quota‐based sampling was used to reflect the population distribution by age‐sex‐province, participants were recruited using non–probability‐based sampling and excluded residents in the territories, limiting generalisability. Participants in the territories are excluded from most national surveys in Canada due to the methodological challenges and costs of conducting population‐based surveys in these remote areas. Second, the survey measures used were self‐reported, which may be subject to response/social desirability bias. Third, this study used a cross‐sectional design, which makes it difficult to determine temporality between the exposure and outcome. To gain a better understanding of the impacts of awareness and belief that alcohol causes cancer on support for alcohol policies, longitudinal research on population‐level interventions that aim to increase awareness and belief is needed. Finally, a prompted measure was used to assess awareness of the alcohol–cancer link, which may overestimate these measures as compared to a more conservative or unprompted measure, such as the use of an open‐ended question that asks participants to list diseases and health conditions that can arise from weekly alcohol consumption [9, 15].

5. Conclusions

This study highlights that believing alcohol causes cancer moderates the relationship between consumer awareness of alcohol as a carcinogen and support for alcohol policies restricting availability and marketing among a large sample of alcohol consumers in Canada. Government‐mandated health warning labels on alcohol containers are a recommended intervention by the WHO [35] for boosting awareness of alcohol‐related health risks, such as increased cancer risk. Longitudinal, real‐world research is needed to test if interventions, such as alcohol warning labels, can raise both consumer awareness and enhance beliefs and acceptance of the alcohol–cancer link.

Author Contributions

Each author certifies that their contribution to this work meets the standards of the International Committee of Medical Journal Editors.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: dar70072‐sup‐0002‐Supinfo.docx.

Weerasinghe A., Forbes S. M., and Hobin E., “Does Believing Alcohol Causes Cancer Moderate the Relationship Between Consumer Awareness of the Alcohol–Cancer Link and Support for Alcohol Policies? Findings From a Canadian Cross‐Sectional Study,” Drug and Alcohol Review 45, no. 1 (2026): e70072, 10.1111/dar.70072.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.