Abstract
Objective
To test whether alcohol labels with a cancer warning and national drinking guidelines are an effective tool for supporting more informed and safer alcohol consumption among drinkers.
Methods
Using a quasi-experimental design, pre-post surveys were conducted with 1647 cohort participants systematically selected in liquor stores in two matched sites in Canada in 2017–2018. Enhanced labels designed according to best practices for effective product labels were applied to alcohol containers in the liquor store in the intervention site for one month, and usual practice continued in the comparison site. Generalized estimating equations tested the differences between sites over time in label salience and processing, and self-reported impact of the labels on drinking behaviours.
Results
After the intervention, recall of the cancer warning label increased to a greater extent in the intervention versus comparison site (adjusted odds ratio (AOR) = 32.2, 95% CI = 5.4, 191.1), but not the national drinking guideline label (AOR = 2.7, 95% CI = 0.2, 31.8). There were significant label effects in the intervention versus comparison site for reading (AOR = 1.8, 95% CI = 1.3, 2.5), thinking about (AOR = 2.0, 95% CI = 1.4, 2.9), and talking with others about (AOR = 2.1, 95% CI = 1.3, 3.6) the labels, as well as self-reported impact to cut down on drinking (AOR = 2.5, 95% CI = 1.3, 4.7) and to drink less (AOR = 2.4, 95% CI = 1.3, 4.3).
Conclusions
Alcohol labels with a cancer warning and national drinking guidelines do a better job conveying risk information and promoting safer consumption than existing practices. Industry has a legal duty to adequately inform consumers about their products and should be mandated to include key information on alcohol containers.
Electronic supplementary material
The online version of this article (10.17269/s41997-020-00320-7) contains supplementary material, which is available to authorized users.
Keywords: Alcohol labels, Alcohol, Policy, Population health, Quasi-experiment, Cancer warning, National alcohol drinking guidelines
Résumé
Objectif
Tester si des étiquettes apposées sur les contenants d’alcool et énonçant les risques de cancer et les lignes directrices nationales sur la consommation d’alcool sont des outils efficaces pour informer les buveurs et les aider à être plus prudents au moment de consommer de l’alcool.
Méthodes
Dans le cadre d’un modèle quasi expérimental, des sondages avant-après ont été menés auprès d’une cohorte de 1 647 participants systématiquement choisis dans des magasins d’alcools de deux endroits ayant des caractéristiques similaires au Canada en 2017-2018. Des étiquettes améliorées conçues conformément aux pratiques exemplaires en matière d’étiquetage efficace ont été apposées sur les contenants d’alcool dans le magasin d’alcools du site d’intervention pendant un mois, tandis que les pratiques habituelles se sont poursuivies dans le site témoin. Des équations d’estimation généralisées ont permis de mesurer, au fil du temps, la différence entre les sites sur le plan de l’importance accordée aux étiquettes et le traitement de l’information y figurant, ainsi que les retombées autodéclarées que ces étiquettes avaient eues sur les habitudes de consommation d’alcool.
Résultats
Après l’intervention, le souvenir de l’étiquette sur les risques de cancer était plus grand au site d’intervention qu’au site témoin [rapport de cotes ajusté (RCA)=32,2, IC 95 % = 5,4, 191,1], mais ce n’était pas le cas pour l’étiquette sur les directives nationales de consommation d’alcool (RCA=2,7, IC 95 % = 0,2, 31,8). Les retombées de l’étiquette étaient significatives au site d’intervention comparativement au site témoin en ce qui concerne la lecture des étiquettes (RCA=1,8, IC 95 % = 1,3, 2,5), la réflexion à propos des étiquettes (RCA=2,0, IC 95 % = 1,4, 2,9), la discussion avec les autres au sujet des étiquettes (RCA=2,1, IC 95 % = 1,3, 3,6) et la réduction autodéclarée de la consommation d’alcool (RCA=2,4, IC 95 % = 1,3, 4,3), ainsi que l’influence perçue de ces étiquettes sur la consommation d’alcool (RCA=2,5, IC 95 % = 1,3, 4,7).
Conclusions
Les étiquettes apposées sur les contenants d’alcool qui énoncent les risques de cancer et les lignes directrices nationales sur la consommation d’alcool sont plus efficaces que les pratiques actuelles pour communiquer les risques associés à la consommation d’alcool et promouvoir une consommation plus sécuritaire. L’industrie est légalement tenue d’informer adéquatement les consommateurs au sujet de ses produits et devrait être obligée d’inclure des renseignements importants sur les contenants d’alcool.
Mots-clés: Étiquettes sur les contenants d’alcool, Alcool, Politique, Santé de la population, Quasi expérimental, Mise en garde de risque de cancer, Lignes directrices nationales sur la consommation d’alcool
Introduction
Alcohol is a leading risk factor for disease and premature mortality (Global Burden of Disease (GBD) 2016 Alcohol Collaborators 2018). It is causally linked to more than 200 health conditions, including seven cancers (GBD 2016 Alcohol Collaborators 2018; Bagnardi et al. 2015). In 2016, cancers were the predominant source of total alcohol-attributable deaths in higher income countries and for ages 50+ (GBD 2016 Alcohol Collaborators 2018). Beyond health effects, alcohol use causes significant social and economic costs to individuals and societies. With global consumption increasing and 2 billion consuming alcohol regularly, the total burden of alcohol in many developed countries, including Canada, exceeds that from all illicit substances combined and is similar to or, by some estimates, greater than that of tobacco (Manthey et al. 2019; Canadian Substance Use Costs and Harms Scientific Working Group 2018).
Despite its burden, awareness of alcohol-related health risks is low, particularly with respect to cancer and national drinking guidelines. Globally, only 13% in some jurisdictions are able to identify alcohol as a cancer risk, while surveys in Canada and England indicate that 26% have heard of national drinking guidelines and 25% know the recommended limits, respectively (Scheideler and Klein 2018; Health Canada 2014; Buykx et al. 2018). Thus, interventions aimed at limiting or promoting moderate alcohol consumption, even when supported by strong evidence, are often resisted by the public and politicians who see alcohol as an integral part of society (Wise 2015).
Supporting safer and informed alcohol use is a critical component of a public health strategy to reduce alcohol-related harms. Alcohol consumption is a complex system-level issue influenced by various interconnected factors (Petticrew et al. 2017). Given this complexity, a comprehensive strategy is required to impact the range of factors driving population-level alcohol consumption and harms. The World Health Organization recommends alcohol labels to raise awareness of and start conversations about alcohol-related health risks (Rekve 2017). Health warnings on tobacco packages are a proven policy for altering consumer knowledge and intentions, as well as contributing to changes in more downstream smoking-related behaviours, such as quit attempts (Hammond 2011). Consistent with this step-wise approach applied in tobacco control, evidence demonstrates the potential for health warnings to weaken public perceptions of alcohol and to increase support for more stringent policies affecting price, availability, and marketing (Al-Hamdani and Smith 2015; Al-Hamdani and Smith 2017; Bates et al. 2018).
Alcohol labels are appealing because of their relatively low cost to regulators, unparalleled reach among drinkers, and higher exposure among the heaviest drinkers (Greenfield 1997). They are unique among alcohol control interventions in that they are delivered at the point of purchase and consumption. Currently, 47 countries mandate alcohol warnings on containers (World Health Organization 2018). However, given the little evidence to inform alcohol label content, format, and size, these elements vary across countries (Martin-Moreno et al. 2013). The common warning statements focus on a specific behaviour or subpopulation, such as drinking when driving or pregnant women, or refer to general harms and excessive drinking. Some countries include label messages stating the number of standard drinks or units of alcohol in a container, or national drinking guidelines. None of the current labels include a cancer warning, except those from South Korea (World Health Organization 2018). Since 2017, alcohol manufacturers in South Korea are able to choose one of three messages, two of which cite cancer risk. Additionally, Ireland passed legislation in late 2018 mandating cancer warnings on alcohol product labels. Yet, because of the pragmatic challenges of evaluating large-scale population-based interventions, evaluations are limited to the inconspicuous US text warnings cautioning about drinking when pregnant or when operating a motor vehicle, and general health harms (Greenfield 1997; Martin-Moreno et al. 2013).
This real-world study experimentally tests the effectiveness of alcohol labels with a cancer warning and national drinking guidelines as a tool for increasing awareness of alcohol-related health risks and promoting safer alcohol consumption. This study investigates the impact of alcohol labels on the following: (i) unprompted recall of label messages, (ii) depth of cognitive processing of label messages, and (iii) self-reported impact on alcohol consumption.
Methods
Intervention labels
The content, format, and design of the labels tested were informed by evidence and best principles for product warnings (Rekve 2017; Hammond 2011; Al-Hamdani and Smith 2015; Greenfield 1997; World Health Organization 2018; Martin-Moreno et al. 2013; Pettigrew et al. 2016), including two studies conducted by our research team (Hobin et al. 2018; Vallance et al. 2018), and two years of consultations with local and international health experts, and community stakeholders. The labels are relatively large in size to make them easily noticed and read, are full colour with a bright yellow background and red border so they stand out on the container, have messages providing new information, and are rotated to avoid wear-out (Fig. 1). Label messages are provided in Canada’s two official languages, English and French.
Fig. 1.
a Intervention alcohol warning labels (actual size 5.0 cm × 3.2 cm). Note: Alcohol containers sold in the liquor store in the intervention site displayed only one of the labels at any one time. b Comparison of labels: Pre-intervention alcohol warning label in the intervention site and the warning label in the comparison site
Study design
A pre-post quasi-experimental study was conducted to test the impact of the intervention labels on all alcohol containers, except local and single-serve beer and cider (~ 3% of products), in one liquor store in the intervention site (Whitehorse, Yukon), relative to two liquor stores in the comparison site (Yellowknife, Northwest Territories) without the intervention labels. These stores were selected to participate as they are the only government monopoly liquor stores in the experimental sites, account for the majority of alcohol sales in these jurisdictions, and are socio-demographically and geographically matched (Statistics Canada 2017; Government of Northwest Territories 2016–2017; Government of Yukon 2016–2017). Additionally, these liquor stores are located in the only two jurisdictions in Canada that have required a post-manufacturer alcohol label on all containers since 1991 that cautions about drinking when pregnant, with an additional warning in Northwest Territories about drinking and operating machinery and general health concerns (Fig. 1) (Government of Northwest Territories 2016–2017; Government of Yukon 2016–2017). Longitudinal surveys were scheduled to assess participant responses in the intervention site four months before and eight months after the intervention labels were implemented, and simultaneously conducted in the comparison site. The intervention labels were launched in the liquor store in the intervention site on November 20th, 2017, with liquor store staff applying the cancer warning and national drinking guidelines labels, and the standard drink label slated to follow shortly thereafter. However, one month into the eight-month intervention period, the government in the intervention site halted their participation in the study and stopped applying labels due to significant pressure from Canada’s national alcohol industry (Austen 2018). Consequently, the follow-up surveys were conducted in the intervention and comparison sites two months after the government halted participation, to capture the impact of the briefer-than-intended intervention. Based on the remaining label stock, approximately 47,000 cancer warning labels and 53,000 national drinking guidelines labels were affixed to alcohol containers before the intervention stopped.
Procedures
At baseline, liquor store customers in the intervention and comparison sites were systematically selected to participate in the survey upon exiting the liquor store using a standard intercept technique of approaching every person that passed a pre-identified landmark in the liquor store. Eligibility for the survey was established through a screening tool assessing age, place of residence, alcohol drinking status, self-reported pregnant or breastfeeding, and purchasing alcohol at the liquor store. Eligible participants completed the survey on a tablet independently without interviewer assistance, and the surveys were conducted simultaneously at the study sites between 10 am and 8 pm by different interviewer teams to avoid time-related effects. Participants received gift cards as remuneration for their time.
At follow-up, participants were sent email invitations and a link to the online survey, plus a monetary incentive. Participants lost to attrition at follow-up were “replenished” using the procedures employed at baseline in liquor stores. Surveys at baseline and follow-up were conducted over 6 weeks and were 18 minutes in length, and the measures were standardized across the two survey waves and sites. Written informed consent was obtained before completing the survey. These procedures were approved by the research ethics boards at Public Health Ontario (ID 2017-010.04) and the University of Victoria (Protocol 17-161).
Sample
Participants in this study were of legal drinking age 19+, residents of either the intervention or comparison site, and, at the time of recruitment; were “current drinkers” (consumed at least one alcoholic drink in the past 30 days); purchased alcohol at the liquor store; and did not self-report being pregnant or breastfeeding.
In total, 1647 unique participants completed at least one of the two surveys, 836 at baseline and 811 at follow-up. According to AAPOR #4, response rates were 9.6% in the intervention and 7.5% in the comparison site (American Association for Public Opinion Research 2011). Overall, 445 (59.0%) participants were retained from baseline to follow-up. Participants lost to follow-up between waves were more likely to have lower education and literacy, and to be in the comparison site.
Measures
Survey items were adapted from indicators used to assess health warning label effectiveness on alcohol and tobacco (Hammond 2011; Hobin et al. 2018; Greenfield 1997; Martin-Moreno et al. 2013; International Agency for Research on Cancer 2008). Unless otherwise stated, baseline measures were anchored with six months prior to baseline, and follow-up measures from the November prior to follow-up.
To assess whether consumers noticed the alcohol labels, participants were asked if they have seen any warning labels on bottles or cans of beer, wine, hard liquor, coolers, or ciders (Yes/No/Do Not Know).
To assess unprompted recall of the cancer warning and drinking guidelines labels, participants were asked to indicate what messages they had seen on warning labels on bottles or cans of beer, wine, or liquor in the relevant look-back period, as open text. When coding the responses, one Research Assistant was blinded to experimental condition, and a second reviewed ambiguous responses to confirm accuracy. Participants who mentioned “cancer” and/or broadly referred to “drinking guidelines” were coded as “Yes” for recall of cancer and of national drinking guidelines separately.
Cognitive processing, the extent to which information is processed or elaborated upon, is the most important determinant of memory and attitude change in response to new information (Cantrell et al. 2013). Depth of processing was assessed using three items, including the frequency with which drinkers attended to the labels (“How often, if at all, have you read or looked closely at the warning labels on bottles and cans of beer, wine, hard liquor, coolers, or ciders?”), and reflected on the labels (“How often, if at all, have you thought about the warning labels on bottles and cans of beer, wine, hard liquor, coolers, or ciders?”). A more distal measure of cognitive processing captured the frequency with which participants talk about the labels with others (“How often have you talked about the warning labels on bottles or cans of beer, wine, hard liquor, coolers, or ciders with others?”). Responses were given on a 5-point Likert scale and dichotomized: 1 = never and 2 = rarely versus 3 = sometimes, 4 = often, and 5 = very often.
To assess the self-reported impact of the labels on drinking, participants were asked, “To what extent, if at all, have warning labels on bottles or cans of beer, wine, hard liquor, coolers, or ciders influenced you to cut down your drinking?” Responses were recorded on a 5-point Likert scale, from “No influence” to “Main influence”, and were dichotomized to indicate at least “some” influence compared with “little” or “no” influence on behaviour. Participants were also asked, “Has the amount of alcohol you are drinking changed as a result of the warning labels on bottles or cans of beer, wine, hard liquor, coolers, or ciders?” Responses included “less” compared with “same amount” and “more”.
Socio-demographics included sex, age, ethnicity, income, education, and health literacy. Ethnicity data were dichotomized (White vs. Other). Annual income was categorized into three levels: low (< $30,000), medium ($30,000–$59,999), and high (≥ $60,000). Education was categorized into three levels: low (completed high school or less), medium (completed trades or college certificate, some university or university certificate below bachelor’s), and high (university degree or post-graduation). Health literacy was assessed using the Newest Vital Sign assessment tool (Weiss et al. 2005), and responses were categorized into three levels: limited (≤ 1 correct response), possibility of limited (2–3 correct responses), and adequate literacy (4–6 correct responses).
Alcohol drinking behaviours were measured using the standard quantity/frequency method (Heeb and Gmel 2005). At each wave, participants were asked how often they drank alcoholic beverages in the past 6 months, and how many drinks they usually drank per occasion. A definition of a “drink” across beverage types was provided in the survey. Responses were combined to provide the mean number of drinks per week and categorized into tertiles based on Canada’s low-risk drinking guidelines: low (≤ 10 for females/15 for males per week), risky (11–19/16–29 per week), high (≥ 20/30 per week) (Butt et al. 2011).
Statistical analyses
To estimate the effect of the labels on recall, depth of processing, and self-reported impact on drinking, seven adjusted logistic generalized estimating equation (GEE) models were used with the logit link applied. GEE models can account for a mix of within-subject correlation that arises from the cohort participants being asked the same questions over multiple survey waves plus the replenishment sample. All models included variables for experimental site, wave, and their interaction. The interaction between site and wave provided a formal test of whether the pattern of change over time in the intervention site was significantly different from that in the comparison site. In each of the models, the following invariant control variables were reported at baseline: sex, ethnicity, and health literacy. Age, education, income, and drinking behaviour measures are time-varying factors reported at each wave. “Prefer Not to Say” and missing responses were removed from all outcome measures. The GEE model estimating recall for the national drinking guidelines label required the addition of a dummy observation to address non-convergence due to a cell count with 0, and health literacy was excluded as the model was overspecified with this covariate. All analyses were conducted using SAS version 9.3 (SAS Institute Inc.; Cary, NC). As per agreement with the local territorial government partners, ethnicity is included in the sample description and adjusted for in the analyses, but not reported in the results.
Results
There were significant differences in the sample characteristics for age, ethnicity, sex, income, alcohol use, and health literacy levels (Table 1).
Table 1.
Sample characteristics at recruitment (initial survey wave for each participant)
| Intervention site (n = 996) | Comparison site (n = 651) | |
|---|---|---|
| n (%) | n (%) | |
| Wave of recruitment | ||
| 1 | 505 (50.7) | 331 (50.8) |
| 2 | 491 (49.3) | 320 (49.2) |
| Age categories*** | ||
| 19–24 | 68 (6.8) | 80 (12.3) |
| 25–44 | 342 (34.3) | 296 (45.5) |
| 45+ | 586 (58.8) | 275 (42.2) |
| Ethnicity*** | ||
| White (vs. Other) | 728 (73.1) | 390 (59.9) |
| Sex** | ||
| Female (vs. male) | 513 (51.5) | 287 (44.1) |
| Education levels* | ||
| Low (completed high school or less) | 205 (20.6) | 149 (22.9) |
| Medium (trades or college certificate, some university or university certificate below bachelor’s) | 352 (35.3) | 233 (35.8) |
| High (bachelor’s degree or higher) | 393 (39.5) | 223 (34.3) |
| Unknown (DK, PNS, missing) | 46 (4.6) | 46 (7.1) |
| Income levels** | ||
| Low (< $30,000) | 152 (15.3) | 76 (11.7) |
| Medium ($30,000 to < $60,000) | 186 (18.7) | 98 (15.1) |
| High (≥ $60,000) | 567 (56.9) | 388 (59.6) |
| Unknown (DK, PNS, missing) | 91 (9.1) | 89 (13.7) |
| Alcohol use levels*** | ||
| Low volume (≤ 10 for females/15 for males per week) | 749 (75.2) | 448 (68.8) |
| Risky volume (11–19/16–29 per week) | 82 (8.2) | 41 (6.3) |
| High volume (≥ 20/30 per week) | 91 (9.1) | 76 (11.7) |
| Unknown (DK, PNS, missing) | 74 (7.4) | 86 (13.2) |
| Health literacy levels*** | ||
| Limited literacy (score ≤ 1) | 305 (30.6) | 233 (35.8) |
| Possibility of limited literacy (score 2–3) | 191 (19.2) | 135 (20.7) |
| Adequate literacy (score 4–6) | 452 (45.4) | 230 (35.3) |
| Unknown (DK, PNS, Missing) | 48 (4.8) | 53 (8.1) |
*Pearson χ2 test, p < 0.05
**Pearson χ2 test, p < 0.01
***Pearson χ2 test, p < 0.001
DK Do Not Know, PNS Prefer Not to Say
Rates of noticing the labels were high at baseline and follow-up in both the intervention (baseline = 80.4%, follow-up = 76.7%) and comparison (baseline = 87.0%, follow-up = 78.5%) sites.
As reported in Table 2 (further details in Supplementary Material—Table A), recall of the cancer warning label increased to a greater extent in the intervention versus the comparison site (+ 24.2% vs. 0.6%, adjusted odds ratio (AOR) = 32.2, 95% CI = 5.4, 191.1) (Fig. 2a) but not the drinking guidelines label (+ 7.3% vs. + 0.7%, AOR = 2.7, 95% CI = 0.2, 31.8) (Fig. 2b). The cancer message was more likely to be recalled by participants with high education (AOR = 1.8, 95% CI = 1.0, 3.1) and adequate literacy (AOR = 1.8, 95% CI = 1.1, 2.9).
Table 2.
Results of generalized estimating equation (GEE) models for label outcomes
| Unprompted recall of the cancer warning | Unprompted recall of drinking guidelines | Reading labels | Thinking about labels | Talking about labels | Intentions to cut back | Self-report of drinking less | |
|---|---|---|---|---|---|---|---|
| AOR (95% CI) | AOR (95% CI) | AOR (95% CI) | AOR (95% CI) | AOR (95% CI) | AOR (95% CI) | AOR (95% CI) | |
| Wave × site | 32.2 (5.4, 191.1)* | 2.7 (0.2, 31.8) | 1.8 (1.3, 2.5)* | 2.0 (1.4, 2.9)* | 2.1 (1.3, 3.6)* | 2.5 (1.3, 4.7)* | 2.4 (1.3, 4.3)* |
| Age | |||||||
| 19–24 | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
| 25–44 | 0.8 (0.4, 1.8) | 0.7 (0.2, 2.2) | 0.9 (0.6, 1.3) | 0.6 (0.4, 0.8)* | 0.7 (0.4, 1.0) | 1.3 (0.7, 2.4) | 1.1 (0.7, 2.0) |
| 45+ | 0.7 (0.3, 1.7) | 0.4 (0.1, 1.2) | 0.7 (0.5, 1.0) | 0.6 (0.4, 0.9)* | 0.6 (0.4, 0.9)* | 0.8 (0.5, 1.6) | 0.9 (0.5, 1.5) |
| Sex | |||||||
| Female (Ref = male) | 1.1 (0.8, 1.5) | 1.0 (0.6, 1.7) | 1.2 (1.0, 1.5)* | 1.3 (1.1, 1.6)* | 1.2 (0.9, 1.5) | 1.5 (1.0, 2.1)* | 1.4 (1.1, 1.9)* |
| Education | |||||||
| Low (completed high school or less) | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
| Medium (trades or college certificate, some university or university certificate below bachelor’s) | 1.0 (0.6, 1.7) | 1.0 (0.4, 2.7) | 1.0 (0.8, 1.4) | 1.1 (0.8, 1.5) | 0.7 (0.5, 1.1) | 1.1 (0.7, 1.7) | 0.8 (0.6, 1.2) |
| High (bachelor’s degree or higher) | 1.8 (1.0, 3.1)* | 2.0 (0.8, 5.3) | 0.9 (0.7, 1.2) | 1.3 (0.9, 1.8) | 1.1 (0.8, 1.6) | 1.1 (0.7, 1.8) | 0.5 (0.3, 0.8)* |
| Unknown | 2.6 (0.9, 7.3) | 0.6 (0.1, 4.2) | 1.3 (0.7, 2.2) | 1.0 (0.6, 1.9) | 1.1 (0.5, 2.3) | 0.7 (0.3, 1.6) | 0.9 (0.5, 1.8) |
| Income | |||||||
| Low (< $30,000) | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
| Medium ($30,000 to < $60,000) | 1.0 (0.5, 2.2) | 1.4 (0.4, 4.5) | 1.0 (0.7, 1.4) | 0.7 (0.5, 1.0)* | 0.7 (0.5, 1.1) | 0.5 (0.3, 0.8)* | 1.2 (0.7, 1.9) |
| High (≥ $60,000) | 1.6 (0.9, 3.0) | 1.5 (0.5, 4.5) | 1.0 (0.7, 1.3) | 0.7 (0.5, 1.0) | 0.7 (0.5, 1.1) | 0.6 (0.4, 1.0)* | 0.9 (0.6, 1.5) |
| Unknown | 1.1 (0.5, 2.4) | 1.0 (0.3, 3.9) | 0.8 (0.5, 1.2) | 0.6 (0.4, 0.9)* | 0.9 (0.6, 1.5) | 0.8 (0.4, 1.3) | 1.0 (0.6, 1.7) |
| Health literacy | |||||||
| Limited literacy (score ≤ 1) | Ref | N/A | Ref | Ref | Ref | Ref | Ref |
| Possibility of limited literacy (score 2–3) | 1.7 (1.0, 2.9) | N/A | 0.9 (0.7, 1.2) | 0.9 (0.6, 1.2) | 1.1 (0.7, 1.5) | 0.5 (0.3, 0.8)* | 1.0 (0.7, 1.5) |
| Adequate literacy (score 4–6) | 1.8 (1.1, 2.9)* | N/A | 1.0 (0.8, 1.3) | 0.8 (0.6, 1.1) | 0.8 (0.6, 1.1) | 0.3 (0.2, 0.4)* | 0.5 (0.4, 0.8)* |
| Unknown | 0.5 (0.1, 1.6) | N/A | 1.0 (0.6, 1.5) | 1.6 (1.0, 2.6) | 1.0 (0.5, 2.0) | 1.0 (0.5, 2.1) | 0.8 (0.4, 1.6) |
| Alcohol use | |||||||
| Low volume (≤ 10 for females/15 for males per week) | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
| Risky volume (11–19/16–29 per week) | 1.1 (0.6, 2.1) | 1.3 (0.5, 2.9) | 0.8 (0.6, 1.1) | 0.9 (0.6, 1.3) | 1.0 (0.7, 1.6) | 1.1 (0.5, 2.1) | 0.9 (0.5, 1.6) |
| High volume (≥ 20/30 per week) | 1.3 (0.7, 2.4) | 0.4 (0.1, 1.7) | 0.7 (0.5, 1.0) | 0.8 (0.6, 1.2) | 1.2 (0.8, 1.8) | 1.0 (0.6, 1.6) | 0.7 (0.4, 1.1) |
| Unknown | 0.6 (0.2, 1.4) | 0.7 (0.2, 3.1) | 0.9 (0.7, 1.3) | 1.0 (0.7, 1.4) | 1.0 (0.6, 1.6) | 1.0 (0.6, 1.6) | 1.0 (0.6, 1.6) |
AOR adjusted odds ratio
*Significant at p < 0.05
Fig. 2.
a–g Impact of alcohol warning labels on label outcomes in the intervention and comparison sites (baseline to follow-up)
Regarding depth of processing, there were significant increases in the intervention relative to the comparison site for reading (+ 5.3% vs. − 8.8%, AOR = 1.8, 95% CI = 1.3, 2.5) (Fig. 2c), thinking about (+ 11.2% vs. − 1.5%, AOR = 2.0, 95% CI = 1.4, 2.9) (Fig. 2d), and talking with others about (+ 11.5% vs. + 1.9%, AOR = 2.0, 95% CI = 1.3, 3.6) the labels (Fig. 2e). Females were more likely to read (AOR = 1.2, 95% CI = 1.0, 1.5) and think about (AOR = 1.3, 95% CI = 1.1, 1.6) label messages versus males (Table 2). Participants aged 25–44 (AOR = 0.6, 95% CI = 0.4, 0.8) or 45+ (OR = 0.6, 95% CI = 0.4, 0.9) and with medium (AOR = 0.7, 95% CI = 0.5, 1.0) or unknown (AOR = 0.6, 95% CI = 0.4, 0.9) income were less likely to think about labels, versus the referent groups. Participants aged 45+ (AOR = 0.6, 95% CI = 0.4, 0.9) were less likely to talk about labels relative to those aged 19–25.
Self-reported influence to cut down on drinking (+ 4.0% vs. − 0.5%, AOR = 2.5, 95% CI = 1.3, 4.7) (Fig. 2f) and reports of drinking less because of the labels (+ 3.7% vs. − 3.3%, AOR = 2.4, 95% CI = 1.3, 4.3) (Fig. 2g) increased to greater extents in the intervention versus the comparison site. Females were more likely to self-report cutting back (AOR = 1.5, 95% CI = 1.0, 2.1) and drinking less (AOR = 1.4, 95% CI = 1.1, 1.9) versus males (Table 2). Participants with medium (AOR = 0.5, 95% CI = 0.3, 0.8) or high (AOR = 0.6, 95% CI = 0.4, 1.0) income and possibility of limited (AOR = 0.5, 95% CI = 0.3, 0.8) or adequate (AOR = 0.3, 95% CI = 0.2, 0.4) health literacy were less likely to report the labels’ influence to cut back on drinking, compared with the referent groups. Participants were less likely to report drinking less due to labels if they had high education (AOR = 0.5, 95% CI = 0.3, 0.8) or adequate literacy (AOR = 0.5, 95% CI = 0.4, 0.8), versus the referent groups.
Discussion
This is the first real-world study to experimentally test the impact of alcohol labels with a cancer warning and national drinking guidelines. After a one-month intervention, significant increases were observed in recall of the cancer warning, label processing, and influence on drinking behaviours. Increases in recall of the drinking guidelines label were not statistically significant, suggesting the label effects on cognitive processing and drinking behaviours were likely driven by the cancer warning. Using alcohol labels to increase consumer awareness of commonly unknown alcohol-related health risks, such as cancer, may help drinkers pay closer attention to their drinking and minimize alcohol-related harms. These findings are consistent with previous research examining the extent to which consumers perceive alcohol-cancer statements and provide support for enhancing alcohol labelling policy (Pettigrew et al. 2016).
The findings also highlight the impact of communicating health information in a novel format. The labels tested are consistent with best practices for effective product labels in that they contain symbols and contrasting colours, are large, rotate, and provide key health information about alcohol of which consumers are mostly unaware (Rekve 2017; Hammond 2011; Al-Hamdani and Smith 2015; Greenfield 1997; Martin-Moreno et al. 2013; Pettigrew et al. 2016). These enhanced labels were more likely to be read, thought about, and talked about with others than the alcohol labels previously used in the intervention site and currently used in the comparison site, both of which were implemented in 1991 (Government of Northwest Territories 2016–2017; Government of Yukon 2016–2017). Studies previously revealed that health warnings on tobacco packages that are new or periodically updated are likely to have greater impact on label salience and processing than older warnings (Hammond 2011). It is conceivable that recall of the national drinking guidelines label was lower because of the complexity of the message (Hammond 2011; Al-Hamdani and Smith 2015; Greenfield 1997; International Agency for Research on Cancer 2008). This label contains three separate messages (daily drinking guidelines for males and females, plus a recommendation for two non-drinking days each week), compared with one message on the cancer warning label. Consistent with evidence of label effectiveness, the visual design of the national drinking guidelines label included a full-colour pictogram of a man and a woman to simplify the information, draw attention to key messages, and support greater cognitive processing (Fig. 1) (Hammond 2011; Martin-Moreno et al. 2013; Levie and Lentz 1982). However, since the intervention was shortened to one month, further research is needed to determine whether repeated exposure to this health information over a longer period may improve its impact.
Labels have the ability to expose drinkers to health messaging on alcohol containers on a consistent basis. However, results of this study suggest receiver characteristics may influence the salience and use of label messages as subgroup differences were observed. Females were significantly more likely to read and think about labels as well as to report that labels influenced them to reduce their drinking. Results also show that older respondents were less likely to think or talk with others about label messages compared with 19–24-year-olds, and participants of higher socio-economic status (SES) were more likely to recall but less likely to act on labels relative to lower SES participants. It is unclear why participants with higher SES were more likely to recall the health messages on labels but less likely to report changing drinking behaviours due to labels; however, one explanation may be that although higher SES is associated with being a current drinker and heavy alcohol consumption, higher SES populations experience fewer adverse alcohol-related health outcomes than lower SES populations (Katikireddi et al. 2017). Overall, these findings show alcohol labels are attended to by vulnerable populations and may be an effective communication tool for targeting those at elevated risks of alcohol harm. Future studies with a larger sample size are needed to test the differential impacts of enhanced alcohol warning labels designed consistent with best practices across key subpopulations and to refine label design to maximize accessibility, such as including picture-based warnings (Hammond 2011; Cantrell et al. 2013).
Limitations
This study has several limitations. First, the intervention was halted one month into the eight-month intervention period with a two-month lag between labelling and the follow-up survey. This brief intervention period and gap in follow-up may have attenuated the labels’ influence and uncertainty remains about their longer-term impact. Next, the study cannot provide representative estimates of the population as participants were recruited from liquor stores in city centres using non-probability-based methods. Finally, this study did not examine the effect of alcohol labels on objective measures of alcohol consumption (e.g., alcohol sales data) and responses may have been subject to self-report bias. Nevertheless, the extent of this bias would be constant across conditions and would not account for differences between experimental conditions over time.
Conclusions
The findings demonstrate that alcohol warning labels do a better job conveying risk information and promoting safer consumption than existing practices. The alcohol industry has opposed warnings on health effects, such as cancer (Austen 2018), which highlights the importance of mandatory alcohol labelling to ensure that consumers are adequately informed.
Electronic supplementary material
(DOCX 19 kb)
Acknowledgements
The authors would like to acknowledge all of our research assistants who helped with data collection, Drs. Mark Petticrew and Melanie Wakefield for their advice and expertise, and the liquor control boards, health and social services, and community partners in Yukon and Northwest Territories for their commitment and support in developing and executing this research.
Funding information
This research was supported by the Health Canada—Substance Use and Addiction Program (grant no. 1718-HQ-000003). The views expressed herein do not necessarily represent the views of Health Canada. TKG was supported by a NIAAA grant (No. P50 AA005595).
Compliance with ethical standards
Conflict of interest
Tim Stockwell received research funds and travel expenses from both the Swedish (Systembolaget) and Finnish (ALKO) government retail alcohol monopolies for the conduct of research into the impacts of their policies on alcohol consumption and related harm, and Thomas K. Greenfield’s research has been partially supported by the National Alcohol Beverage Control Association (NABCA). All other authors declare no conflict of interest.
Ethical standards disclosure
This study was conducted according to the guidelines in the Declaration of Helsinki and procedures involving human subjects were approved by the ethics review boards at Public Health Ontario and University of Victoria.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- Al-Hamdani M, Smith SM. Alcohol warning label perceptions: emerging evidence for alcohol policy. Can J Public Health. 2015;106(6):e395–e400. doi: 10.17269/CJPH.106.5116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Al-Hamdani M, Smith SM. Alcohol warning label perceptions: do warning sizes and plain packaging matter? J Stud Alcohol Drugs. 2017;78(1):79–87. doi: 10.15288/jsad.2017.78.79. [DOI] [PubMed] [Google Scholar]
- American Association for Public Opinion Research (2011). Standard definitions: final dispositions of case codes and outcomes for surveys. 7th ed. https://www.esomar.org/uploads/public/knowledge-and-standards/codes-and-guidelines/ESOMAR_Standard-Definitions-Final-Dispositions-of-Case-Codes-and-Outcome-Rates-for-Surveys.pdf. Accessed December 7, 2018.
- Austen, I. (2018). Yukon government gives in to liquor industry on warning label experiment. New York Times https://www.nytimes.com/2018/01/06/world/canada/yukon-liquor-alcohol-warnings.html. Accessed September 6, 2018.
- Bagnardi V, Rota M, Botteri E, et al. Alcohol consumption and site-specific cancer risk: a comprehensive dose–response meta-analysis. Br J Cancer. 2015;112(3):580–593. doi: 10.1038/bjc.2014.579. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bates S, Holmes J, Gavens L, et al. Awareness of alcohol as a risk factor for cancer is associated with public support for alcohol policies. BMC Public Health. 2018;18:688. doi: 10.1186/s12889-0180558108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Butt, P., Beirness, D., Gliksman, L., Paradis, C., Stockwell, T. (2011). Alcohol and health in Canada: a summary of evidence and guidelines for low risk drinking. Ottawa: Canadian Centre on Substance Abuse, 2011. http://www.ccdus.ca/Resource%20Library/2011-Summary-of-Evidence-and-Guidelines-for-Low-Risk%20Drinking-en.pdf. Accessed December 7, 2018.
- Buykx P, Li J, Gavens L, Hooper L, Gomes de Matos E, Holmes J. Self-reported knowledge, correct knowledge, and use of UK drinking guidelines among a representative sample of the English population. Alcohol Alcohol. 2018;53(4):453–460. doi: 10.1093/alcalc/agx127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Canadian Substance Use Costs and Harms Scientific Working Group (2018). Canadian Institute for Substance Use Research; Canadian Centre on Substance Use and Addiction. Canadian substance use costs and harms (2007–2014). Ottawa, ON: Canadian Centre on Substance Use and Addiction; 2018. http://www.ccsa.ca/Resource%20Library/CSUCH-Canadian-Substance-Use-Costs-Harms-Report-2018-en.pdf. .
- Cantrell J, Vallone DM, Thrasher JF, et al. Impact of tobacco-related health warning labels across socioeconomic, race, and ethnic groups: results from a randomized web-based experiment. PLoS One. 2013;8(1):e52206. doi: 10.1371/journal.pone.0052206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Global Burden of Disease 2016 Alcohol Collaborators Alcohol use and burden for 195 countries and territories, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet. 2018;392(10152):1015–1035. doi: 10.1016/S0140-6736(18)31310-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Government of Northwest Territories (2016-2017). Northwest Territories Liquor Commission 2016–2017 63rd Annual Report. Available at: http://library.assembly.gov.nt.ca/2017/F/a373811_td_451-182_Northwest_Territories_Liquor_Commission_63rd_annual_report.pdf.
- Government of Yukon (2016-2017). Yukon Liquor Corporation Annual Report 2016–2017. Available at: http://www.ylc.yk.ca/pdf/YLC_annual_report_2016_17_WEB.pdf.
- Greenfield T. Warning labels: evidence on harm reduction from long-term American surveys. In: Plant M, Single E, Stockwell T, editors. Alcohol: minimizing the harm. London: Free Association Books; 1997. pp. 105–125. [Google Scholar]
- Hammond D. Health warning messages on tobacco products: a review. Tob Control. 2011;20(5):327–337. doi: 10.1136/tc.2010.037630. [DOI] [PubMed] [Google Scholar]
- Health Canada (2014). Canadian alcohol and drug use monitoring survey: summary of results for 2012. Government of Canada Web site. https://www.canada.ca/en/health-canada/services/health-concerns/drug-prevention-treatment/drug-alcohol-use-statistics/canadian-alcohol-drug-use-monitoring-survey-summary-results-2012.html#s7. Accessed November 27, 2018.
- Heeb JL, Gmel G. Measuring alcohol consumption: a comparison of graduated frequency, quantity frequency, and weekly recall diary methods in a general population survey. Addict Behav. 2005;30(3):403–413. doi: 10.1016/j.addbeh.2004.04.022. [DOI] [PubMed] [Google Scholar]
- Hobin E, Vallance K, Zuo F, et al. Testing the efficacy of alcohol labels with standard drink information and national drinking guidelines on consumers’ ability to estimate alcohol consumption. Alcohol Alcohol. 2018;53(1):3–11. doi: 10.1093/alcalc/agx052. [DOI] [PubMed] [Google Scholar]
- International Agency for Research on Cancer (2008); World Health Organization. IARC handbooks of cancer prevention: methods for evaluating tobacco control policies. Vol 12. Lyon, France: International Agency for Research on Cancer; 2008.
- Katikireddi SV, Whitley E, Lewsey J, Gray L, Leyland AH. Socioeconomic status as an effect modifier of alcohol consumption and harm: analysis of linked cohort data. Lancet Public Health. 2017;2:e267–ee76. doi: 10.1016/S2468-2667(17)30078-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Levie WH, Lentz R. Effects of text illustrations: a review of research. Educ Technol Res Dev. 1982;30(4):193–232. [Google Scholar]
- Manthey J, Shield KD, Rylett M, Hasan OSM, Probst C, Rehm J. Global alcohol exposure between 1990 and 2017 and forecasts until 2030: a modelling study. Lancet. 2019;393:2493–2502. doi: 10.1016/S0140-6736(18)32744-2. [DOI] [PubMed] [Google Scholar]
- Martin-Moreno JM, Harris ME, Breda J, Møller L, Alfonso-Sanchez JL, Gorgojo L. Enhanced labelling on alcoholic drinks: reviewing the evidence to guide alcohol policy. Eur J Pub Health. 2013;23(6):1082–1087. doi: 10.1093/eurpub/ckt046. [DOI] [PubMed] [Google Scholar]
- Petticrew, M., Shemilt, I., Lorenc, T., et al. (2017). Alcohol advertising and public health: systems perspectives versus narrow perspectives. J Epidemiol Community Health, 71(3). 10.1136/jech-2016-207644. [DOI] [PubMed]
- Pettigrew S, Jongenelis MI, Glance D, et al. The effect of cancer warning statements on alcohol consumption intentions. Health Educ Res. 2016;31(1):60–69. doi: 10.1093/her/cyv067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rekve, D. (2017). World Health Organization. An update from the World Health Organization with a focus on labelling https://eceuropaeu/health/sites/health/files/alcohol/docs/ev_20171107_co09b_enpdf Published November 15, 2017. Accessed December 6, 2018.
- Scheideler JK, Klein WMP. Awareness of the link between alcohol consumption and cancer across the world: a review. Cancer Epidemiol Biomark Prev. 2018;27(4):429–437. doi: 10.1158/1055-9965.EPI-17-0645. [DOI] [PubMed] [Google Scholar]
- Statistics Canada (2017). Census profile, 2016 census. Catalogue no. 98-316-X2016001. Statistics Canada web site. https://www12.statcan.gc.ca/census-recensement/2016/dp-pd/prof/index.cfm?Lang=E. Released February 8, 2017. Updated May 30, 2018. Accessed December 7, 2018.
- Vallance K, Romanovska I, Stockwell T, Hammond D, Rosella L, Hobin E. “We have a right to know”: exploring consumer opinions on content, design, and acceptability of enhanced alcohol labels. Alcohol Alcohol. 2018;53(1):20–25. doi: 10.1093/alcalc/agx068. [DOI] [PubMed] [Google Scholar]
- Weiss B, Mays M, Martz W, et al. Quick assessment of literacy in primary care: the newest vital sign. Ann Fam Med. 2005;3(6):514–522. doi: 10.1370/afm.405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wise J. Responsibility deal for alcohol has failed, says report. BMJ. 2015;351:h6026. doi: 10.1136/bmj.h6026. [DOI] [PubMed] [Google Scholar]
- World Health Organization (2018). Global Information System on Alcohol and Health (GISAH). Global Health Observatory data repository: health warnings on alcohol containers by country. World Health Organization Web site http://appswhoint/gho/data/viewmain55920 Updated September 4, 2018.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
(DOCX 19 kb)