Quantifying the Nature and Extent of Children's Real-time Exposure to Alcohol Marketing in Their Everyday Lives Using Wearable Cameras: Children's Exposure via a Range of Media in a Range of Key Places

Public health and harmful advertising: The nature and extent of children's real-time exposure to unhealthy commodity marketing

The consumption of unhealthy commodities (UHCs) - including unhealthy food, alcohol, tobacco and gambling products - contributes to substantial public health harm and significant social and economic costs. Consumption of UHCs is driven, in part, by increasingly sophisticated, persuasive and extensive product marketing. In Aotearoa New Zealand (NZ), UHC marketing is largely self-regulated by industry but varies by individual commodity. This paper aims to identify the nature and extent of children's exposure to UHC marketing in NZ and consider policy implications. Using data collected from wearable cameras, children's brand marketing exposures were analysed to identify their rate of exposure to UHC marketing, as well as 'Healthy' (core food/social marketing) and 'Other' marketing. Children (N = 90) were exposed to UHC marketing on average 76.2 times per day, nearly 2.5 times their daily exposure to 'Healthy' marketing. Children were exposed to UHC marketing through a more diverse range of settings and mediums than core food and social marketing. Over half of children's unhealthy food (54.9 %) and alcohol (51.9 %) marketing exposures were attributed to multinational corporations (MNCs). Children's exposure to each UHC category generally aligned with the level of regulation over that commodity in NZ. Overall, these findings support comprehensive statutory marketing regulation over UHCs, both in NZ and likely internationally given the global nature of UHCs. Statutory marketing regulation would reduce children's exposure to UHC marketing and protect public health. Given the extensive similarities among UHCs, policymakers should consider a joint regulatory approach.

Public and professional stakeholders' perceptions of alcohol advertising and availability policies: A qualitative study

INTRODUCTION

Reducing alcohol availability and restricting alcohol advertising are effective ways to reduce harm from alcohol. Implementation of public health policies involves collaboration between different stakeholders, and is influenced by public opinion. This paper explores public and professional stakeholders' perceptions of alcohol advertising and availability policies. It is the first to capture consensus and divergence in narratives of these stakeholders.

METHODS

We conducted semi-structured interviews with 14 stakeholders from third sector organisations, government, public health and alcohol licensing in Scotland. We conducted 11 online focus groups with 45 participants, living in neighbourhoods in Scotland characterised by varying levels of urbanity, deprivation and retail density change. We gave participants a list of policies and discussed their views on acceptability, feasibility and likely success.

RESULTS

Despite general consensus that regulation of alcohol advertising is an important priority, public stakeholders were concerned about the feasibility of advertising interventions and potential unintended consequences. While professional stakeholders were in favour of regulating alcohol availability, public stakeholders had misgivings about feasibility and effectiveness. When prompted to discuss specific interventions, similar views about protecting children and achieving cultural change emerged.

DISCUSSION AND CONCLUSIONS

This study highlights the importance of policy makers and other stakeholders to consider public stakeholders' opinions on alcohol policy and understanding that their views may be influenced by competing framings of alcohol problems. Attempts to increase support for alcohol control policies need to consider people's concerns about the effectiveness and potential unintended consequences of these policies, and the wider social context of alcohol consumption.

The estimated health impact of alcohol interventions in New Zealand: A modelling study

AIMS

To estimate the health impacts of key modelled alcohol interventions among Māori (indigenous peoples) and non-Māori in New Zealand (NZ).

DESIGN

Multi-stage life-table intervention modelling study. We modelled two scenarios: (1) business-as-usual (BAU); and (2) an intervention package scenario that included a 50% alcohol tax increase, outlet density reduction from 63 to five outlets per 100 000 people, outlet hours reduction from 112 to 50 per week and a complete ban on all forms of alcohol marketing.

SETTING AND PARTICIPANTS

The model's population replicates the 2018 NZ population by ethnicity (Māori/non-Māori), age and sex.

MEASUREMENTS

Alcohol consumption was estimated using nationally representative survey data combined with sales data and corrected for tourist and unrecorded consumption. Disease incidence, prevalence and mortality were calculated using Ministry of Health data. We used dose-response relationships between alcohol and illness from the 2016 Global Burden of Disease study and calculated disability rates for each illness. Changes in consumption were based on the following effect sizes: total intervention package [-30.3%, standard deviation (SD) = 0.02); tax (-7.60%, SD = 0.01); outlet density (-8.64%, SD = 0.01); outlet hours (-9.24%, SD = 0.01); and marketing (-8.98%, SD = 0.02). We measured health gain using health-adjusted life years (HALYs) and life expectancy.

FINDINGS

Compared with the BAU scenario, the total alcohol intervention package resulted in 726 000 [95% uncertainty interval (UI) = 492 000-913 000] HALYs gained during the life-time of the modelled population. Māori experienced greater HALY gains compared with non-Māori (0.21, 95% UI = 0.14-0.26 and 0.16, 95% UI = 0.11-0.20, respectively). When modelled individually, each alcohol intervention within the intervention package produced similar health gains (~200 000 HALYs per intervention) owing to the similar effect sizes.

CONCLUSIONS

Modelled interventions for increased alcohol tax, reduced availability of alcohol and a ban on alcohol marketing among Māori and non-Māori in New Zealand (NZ) suggest substantial population-wide health gains and reduced health inequities between Māori and non-Māori.

Media coverage of commercial industry activities impacting Aboriginal and Torres Strait Islander health, 2018-2022

This study aimed to examine the extent and nature of Australian news media coverage of commercial industry activities that explicitly speak to Aboriginal and Torres Strait Islander contexts. We undertook content and framing analysis of Australian newspaper and online media articles published between January 2018 and March 2022 that included terms related to 'Indigenous', 'commercial' and 'health'. Analysis focused on the nature of coverage, framing of responsibility, patterns over time and stakeholder representation. Forty-six media articles were included in the analysis. Half of these articles related to the actions of three companies (Woolworths, WAM Clothing and Rio Tinto). Most articles described negative health and well-being impacts of commercial activity, while four described positive impacts. The most common voice represented in media articles was from industry (n = 25). Aboriginal or Torres Strait Islander voices were represented in 21 articles. This analysis highlights how commercial activities in Australia are reported to negatively influence Aboriginal and Torres Strait Islander health and well-being, and that industry voices are more commonly represented in the media related to these issues.

Under the influence: system-level effects of alcohol industry-funded health information organizations

There is now an established body of evidence that the alcohol industry seeks to obstruct public health policies that could affect the availability, affordability or marketing of alcohol. In parallel, the alcohol industry is active in funding corporate social responsibility initiatives, with a particular focus on 'responsible drinking' campaigns, often facilitated by national-level charities established and/or funded by the alcohol industry and associated organizations. While evidence continues to grow regarding biases in the content produced by such health information organizations, they remain active in partnerships with government health departments on national health promotion campaigns and provide a range of health-related information to the public, community organizations and schools. To understand the implications of such access for policymakers, researchers and the public, there is a need to consider the wider, system-level influences of such organizations and their place in wider alcohol industry strategies. In this article, we describe evolving evidence of the direct and indirect strategic effects of such organizations and demonstrate how they serve key roles in the alcohol industry through their existence, content, partnerships and public profiles. We end by considering the implications for how we conceptualize charities established or funded (entirely or partly) by harmful commodity industries, and to what extent current conflicts of interest guidelines are sufficiently effective.

The nature and extent of outdoor alcohol marketing in Wellington, New Zealand: a longitudinal audit and spatio-temporal analysis of outdoor alcohol marketing

AIM

exposure to alcohol marketing is associated with increased consumption. We aimed to (i) measure the nature and extent of outdoor alcohol marketing within a high-density urban neighbourhood and (ii) examine temporal and spatial trends in alcohol marketing.

METHODS

this study used a longitudinal design to monitor paid advertising in public spaces over two 10-week periods in Wellington, New Zealand (Nov-Jan 2020-2021, Nov-Jan 2021-2022). The data were collected on-foot following an established route once a week using a phone camera, which also recorded gps data of ad locations. Temporal and spatial trends in alcohol ad prevalence were assessed.

RESULTS

over the study period, 13% (n = 1619) of all ads (n = 12,472) were for alcohol. Alcohol ads were predominately for spirits (29%), ready-to-drink (27%) and beer (23%). Almost half of all alcohol ads (49%) did not contain a responsible consumption message, while those with a message were de-emphasized relative to promotional features. A temporal trend was observed in 2020, whereby alcohol marketing decreased over the summer, but this trend was not reflected in 2021. Alcohol ads were more likely than non-alcohol ads to be placed in premium positions on roads of high pedestrian and motor vehicle traffic.

CONCLUSION

alcohol marketing is common in urban centres. Local and central government policy could substantially reduce the levels of alcohol marketing exposure via outdoor marketing.

Exposure to e-cigarette advertising and young people's use of e-cigarettes: A four-country study

INTRODUCTION

The World Health Organization recommends banning all forms of e-cigarette advertising, promotion, and sponsorship. The aims of the present study were to: 1) examine young people's exposure to e-cigarette advertising across a wide range of media in four diverse countries; and 2) identify any association between the number of different types of media exposures and e-cigarette use.

METHODS

A cross-sectional online survey was administered to approximately 1000 people aged 15-30 years in Australia, China, India, and the United Kingdom (n=4107). The survey assessed demographic characteristics, e-cigarette and tobacco use, numbers of friends and family members who vape, and exposure to multiple forms of e-cigarette advertising (e.g. television, radio, print, and various types of social media). Descriptive analyses were conducted on those who had heard of e-cigarettes (n=3095, significance threshold p<0.001) and a logistic regression analysis was used to identify factors associated with e-cigarette ever use (significance threshold p<0.05).

RESULTS

The majority (85%) of respondents who had heard of e-cigarettes reported being exposed to e-cigarette advertising on at least one type of media, and the average number of types of media to which respondents were exposed was 5 (range: 0-17). The number of media types was significantly associated with ever use of e-cigarettes (OR=1.05; 95% CI: 1.02-1.08, p=0.001).

CONCLUSIONS

Despite advertising restrictions in place in all four countries, large majorities of young people reported being exposed to e-cigarette advertising. Social media and advertising on/around vape shops and other retailers appear to be key exposure locations. Urgent attention is needed to address these forms of exposure given their apparent association with e-cigarette use.

The promise of zero-shot learning for alcohol image detection: comparison with a task-specific deep learning algorithm

Exposure to alcohol content in media increases alcohol consumption and related harm. With exponential growth of media content, it is important to use algorithms to automatically detect and quantify alcohol exposure. Foundation models such as Contrastive Language-Image Pretraining (CLIP) can detect alcohol exposure through Zero-Shot Learning (ZSL) without any additional training. In this paper, we evaluated the ZSL performance of CLIP against a supervised algorithm called Alcoholic Beverage Identification Deep Learning Algorithm Version-2 (ABIDLA2), which is specifically trained to recognise alcoholic beverages in images, across three tasks. We found ZSL achieved similar performance compared to ABIDLA2 in two out of three tasks. However, ABIDLA2 outperformed ZSL in a fine-grained classification task in which determining subtle differences among alcoholic beverages (including containers) are essential. We also found that phrase engineering is essential for improving the performance of ZSL. To conclude, like ABIDLA2, ZSL with little phrase engineering can achieve promising performance in identifying alcohol exposure in images. This makes it easier for researchers, with little or no programming background, to implement ZSL effectively to obtain insightful analytics from digital media. Such analytics can assist researchers and policy makers to propose regulations that can prevent alcohol exposure and eventually prevent alcohol consumption.

The commercial determinants of Indigenous health and well-being: a systematic scoping review

INTRODUCTION

Health inequity within Indigenous populations is widespread and underpinned by colonialism, dispossession and oppression. Social and cultural determinants of Indigenous health and well-being are well described. Despite emerging literature on the commercial determinants of health, the health and well-being impacts of commercial activities for Indigenous populations is not well understood. We aimed to identify, map and synthesise the available evidence on the commercial determinants of Indigenous health and well-being.

METHODS

Five academic databases (MEDLINE Complete, Global Health APAPsycInfo, Environment Complete and Business Source Complete) and grey literature (Australian Indigenous HealthInfoNet, Google Scholar, Google) were systematically searched for articles describing commercial industry activities that may influence health and well-being for Indigenous peoples in high-income countries. Data were extracted by Indigenous and non-Indigenous researchers and narratively synthesised.

RESULTS

56 articles from the USA, Canada, Australia, New Zealand, Norway and Sweden were included, 11 of which were editorials/commentaries. The activities of the extractive (mining), tobacco, food and beverage, pharmaceutical, alcohol and gambling industries were reported to impact Indigenous populations. Forty-six articles reported health-harming commercial practices, including exploitation of Indigenous land, marketing, lobbying and corporate social responsibility activities. Eight articles reported positive commercial industry activities that may reinforce cultural expression, cultural continuity and Indigenous self-determination. Few articles reported Indigenous involvement across the study design and implementation.

CONCLUSION

Commercial industry activities contribute to health and well-being outcomes of Indigenous populations. Actions to reduce the harmful impacts of commercial activities on Indigenous health and well-being and future empirical research on the commercial determinants of Indigenous health, should be Indigenous led or designed in collaboration with Indigenous peoples.

The distribution of harmful product marketing in public outdoor spaces and the effectiveness of marketing bans

Children's exposure to the marketing of harmful products in public outdoor spaces may influence their consumption of those products and affect health into adulthood. This study aimed to: i) examine the spatial distribution of children's exposure to three types of marketing-related 'harms' (alcohol, unhealthy food, and gambling) in outdoor spaces in the Wellington region, New Zealand/Aotearoa; ii) compare differences in the distribution of harms by socioeconomic deprivation; and iii) estimate the effectiveness of different policies that ban such marketing. Data were from 122 children aged 11-13y who wore wearable cameras and GPS devices for four consecutive days from July 2014 to June 2015. Images were analysed to identify harmful product marketing exposures in public outdoor spaces. Eight policy scenarios were examined to identify the effectiveness of marketing bans, for all children and by socioeconomic deprivation. Children's ratio of harmful marketing was higher for children from high deprivation households and was also found to cluster, with hots spots observed around city centers. The effectiveness of marketing bans depended on the target setting and ban area, with banning 400 m around bus stops leading to the largest reduction. Effectiveness varied also by type of harm and socioeconomic deprivation. For example, banning alcohol marketing in residential areas was estimated to have a larger effect on exposure reduction for children from high deprivation households. Our findings suggest that alcohol, unhealthy food and gambling marketing often cluster outdoors and that targeted bans of such marketing would likely improve child health and, for some banning scenarios, promote equity.

Long-term trends in adolescent alcohol, tobacco and cannabis use and emerging substance use issues in Aotearoa New Zealand

This narrative review summarises the latest evidence on the causes and consequences of substance use in adolescence and describes long-term trends in adolescent alcohol, tobacco and cannabis use in Aotearoa. Adolescence is a time of rapid brain development when young people are uniquely vulnerable to the risks of substance use. It is a major cause of health and social harm in this age group and can affect adult outcomes and the health of the next generation. Therefore, substance use trends are central to understanding the current and future state of child and youth wellbeing in Aotearoa. Adolescent use of alcohol, tobacco and cannabis peaked in the late 1990s/early 2000s, then declined rapidly, and prevalence is now much lower than 20 years ago. However, levels of adolescent binge drinking remain high by international standards and disparities in tobacco and cannabis use by ethnicity and socioeconomic status are wide. Evidence suggests we may again be at a turning point, with-long term declines stalling or reversing in the past 2-5 years, and vaping emerging as a new risk. Greater investment in primary prevention is indicated, including restrictions on alcohol marketing and availability, and alleviation of poverty, racism and marginalisation.

An objective assessment of children's exposure to brand marketing in New Zealand (Kids'Cam): a cross-sectional study

BACKGROUND

Marketing promotes values of consumerism and overconsumption, and negatively affects children's wellbeing and psychological development. The threat marketing poses to planetary health is just being realised. However, little is known about children's exposure to marketing at an aggregate level. Using an objective method of wearable cameras, we aimed to determine the nature and extent of children's exposure to marketing.

METHODS

Kids'Cam was a cross-sectional study of children aged 11-13 years in New Zealand, from which we randomly selected a sample of 90 children. Children wore cameras from when they woke up until they went to sleep for four consecutive days (Thursday-Sunday) that captured images at an angle of 136° every 7 s for exposure to marketing. Marketing brands were categorised into three groups: core food and social marketing messages, harmful commodities (eg, non-core food, alcohol, and gambling), or other. Exposure rates by marketing medium, setting, and product category were calculated using negative binomial regression models.

FINDINGS

From June 21, 2014, to June 30, 2015, we recruited 168 children, and randomly selected data from 90 children for the present study. Children in this study were exposed to a mean of 554 brands per 10 h day (95% CI 491-625), nearly a brand a minute, through multiple mediums (predominantly brand labels [36% of exposures] and product packaging [22%]) and mostly in schools (43%) and at home (30%). Food and beverages (20% of exposures) were the dominant product category. The most pervasive marketing brands typically sold a range of products across more than one product category (eg, children were exposed to Nike on average 20 exposures per day). Children were exposed to more than twice as many harmful commodities (mean 76 per 10 h day [95% CI 55-105]) as core food and social marketing messages (32 [26-39]) per day.

INTERPRETATION

We found that children are repeatedly exposed to marketing through multiple mediums and across all settings, and our findings suggests that marketing privileges particular messages, for example, marketing of harmful commodities. Given the key role marketing plays in establishing and supporting consumption norms, and perpetuating the normalisation of overconsumption which contributes to environmental degradation, these findings suggest an urgent need to reduce marketing to promote planetary health.

FUNDING

This research was funded by a University of Otago, Dean's Research Grant. The Kids'Cam study was funded by a Health Research Council of New Zealand Programme Grant (13/724).

Using Wearable Cameras to Investigate Health-Related Daily Life Experiences: A Literature Review of Precautions and Risks in Empirical Studies

Automated, wearable cameras can benefit health-related research by capturing accurate and objective information about individuals' daily experiences. However, wearable cameras present unique privacy- and confidentiality-related risks due to the possibility of the images capturing identifying or sensitive information from participants and third parties. Although best practice guidelines for ethical research with wearable cameras have been published, limited information exists on the risks of studies using wearable cameras. The aim of this literature review was to survey risks related to using wearable cameras, and precautions taken to reduce those risks, as reported in empirical research. Forty-five publications, comprising 36 independent studies, were reviewed, and findings revealed that participants' primary concerns with using wearable cameras included physical inconvenience and discomfort in certain situations (e.g., public settings). None of the studies reviewed reported any serious adverse events. Although it is possible that reported findings do not include all risks experienced by participants in research with wearable cameras, our findings suggest a low level of risk to participants. However, it is important that investigators adopt recommended precautions, which can promote autonomy and reduce risks, including participant discomfort.

Awareness of alcohol marketing one year after initial implementation of Ireland's Public Health (Alcohol) Act and during the COVID-19 pandemic

BACKGROUND

The Republic of Ireland is introducing new controls on alcohol marketing, starting in November 2019 with restrictions on some outdoor and cinema advertising, and a ban on public transport advertising. We examined changes in marketing awareness one year after initial implementation and during the COVID-19 pandemic.

METHODS

Repeat online cross-sectional surveys with adults in Ireland conducted October 2019 (n = 1,007) and October 2020 (n = 1,020). Participants self-reported past-month awareness of alcohol marketing and completed the Alcohol Use Disorders Identification Test-Concise (AUDIT-C). Current drinkers were categorised as those reporting heavy episodic drinking at least monthly and higher-risk drinkers (≥5 AUDIT-C).

RESULTS

In both waves, most participants recalled some marketing awareness (94.1% vs. 93.8%). For 9/13 activities measured in both waves, there were decreases in the proportion reporting any awareness and frequency of awareness, including for the newly restricted activities. For example, any awareness of public transport advertising decreased between waves (ORAdj = 0.66, 95%CI: 0.53-0.81). In both waves, higher past-month awareness was associated with at least monthly heavy episodic drinking and higher-risk consumption.

CONCLUSION

We recommend a precautionary interpretation. It is plausible that both Ireland's initial controls and COVID-19 restrictions contributed to decreases in awareness, but longer-term evaluation is required to determine relative contribution.

The Frequency and Context of Snacking among Children: An Objective Analysis Using Wearable Cameras

Snacking is a common eating behaviour, but there is little objective data about children’s snacking. We aimed to determine the frequency and context of children’s snacking (n = 158; mean age = 12.6 years) by ethnicity, gender, socioeconomic deprivation and body mass index (BMI) children. Participants wore wearable cameras that passively captured images of their surroundings every seven seconds. Images (n = 739,162) were coded for snacking episodes, defined as eating occasions in between main meals. Contextual factors analysed included: snacking location, food source, timing, social contact and screen use. Rates of total, discretionary (not recommended for consumption) and healthful (recommended for consumption) snacking were calculated using negative binomial regression. On average, children consumed 8.2 (95%CI 7.4, 9.1) snacks per day, of which 5.2 (95%CI 4.6, 5.9) were discretionary foods/beverages. Children consumed more discretionary snacks than healthful snacks in each setting and at all times, including 15.0× more discretionary snacks in public spaces and 2.4× more discretionary snacks in schools. Most snacks (68.9%) were sourced from home. Girls consumed more total, discretionary and healthful snacks than boys, and Māori and Pacific consumed fewer healthful snacks than New Zealand (NZ) Europeans. Results show that children snack frequently, and that most snacking involves discretionary food items. Our findings suggest targeting home buying behaviour and environmental changes to support healthy snacking choices.

An objective methodology capturing online commodity marketing and other harms

Increasingly life is lived online, yet little is known about the actual nature and extent of online content that people view due to the difficulty of recording real time exposure. This includes people's exposure to harmful commodity marketing. This study aimed to develop a methodology to assess the nature and extent of exposure to, and engagement with, unhealthy commodity marketing and other public health harms online, particularly children's exposure. A convenience sample of 16 young adult participants (aged 21-29) recorded their device usage for 2 days using Zoom software. Data were coded and analysed to assess the nature and extent of marketing for alcohol, gambling, junk food and smoking products. Four focus groups were conducted with participants to explore their data collection and coding experiences, and results assessed using thematic analysis. The study found that, with some modifications, this method was feasible for gathering real-time objective data from the online world that can be analysed for a range of public health harms, including marketing of unhealthy commodities. Larger studies are recommended to build global evidence for public health action in the online world.

Depictions of Alcohol in Australian TV 'Bachelor In Paradise': A Content Analysis

AIMS

To quantify the depictions of alcohol in the popular Australian reality TV show-Bachelor in Paradise (season 1: 2018).

METHODS

All 16 episodes were coded in 1-min intervals for the presence of alcoholic beverage related content and non-alcoholic beverage content, and the categories of actual use, implied use and other references.

RESULTS

Alcohol was highly prevalent in all episodes. Alcohol content occurred frequently, with 70.7% of intervals having any alcohol content. Actual alcohol use occurred in 31.9% of 1-min intervals, implied alcohol use occurred in 63.4% of intervals and other alcohol references occurred in 14.0% of intervals. Alcohol content was present in the first or second 1-min interval of all 16 episodes. Alcohol content was more than twice as prevalent as non-alcoholic content (34.0%).

CONCLUSIONS

The high volume of alcohol content depicted in the show is of concern, due to the important influence it may have on the audience. Vulnerable viewers, especially minors and young adults, are being exposed to ubiquitous alcohol references. This may influence their perceptions of normal alcohol use, their attitudes toward alcohol and their own consumption of alcohol. A stronger regulatory regime is required in Australia to protect young people more effectively from depictions in television programs.

Structural racism in the built environment: Segregation and the overconcentration of alcohol outlets

Structural racism, evidenced in practices like residential racial segregation, has been linked to health inequities. We examined the relationship between an adverse environmental factor (alcohol outlet overconcentration), segregated neighborhoods, and county alcohol policy in Louisiana and Alabama to investigate this link. Multilevel analysis revealed high outlet density associated with segregated counties and predominantly black census tracts in counties with restrictive alcohol policy. This inverse association between policies designed to limit alcohol availability and overconcentration of outlets in black neighborhoods warrants consideration by policymakers given links between outlet density and health inequities. Consideration of these findings in historical context suggests these policies may function as a contemporary actualization of the historical use of alcohol policy to subjugate black people in the South, now over-concentrating instead of prohibiting access.

The Frequency and Healthfulness of Food and Beverage Advertising in Movie Theatres: A Pilot Study Conducted in the United States and Canada

The marketing of unhealthy foods and beverages contributes to childhood obesity. In Canada and the United States, these promotions are self-regulated by industry. However, these regulations do not apply to movie theatres, which are frequently visited by children. This pilot study examined the frequency and healthfulness of food advertising in movie theatres in the United States and Canada. A convenience sample of seven movie theatres in both Virginia (US) and Ontario (Canada) were visited once per month for a four-month period. Each month, ads in the movie theatre environment and before the screening of children's movies were assessed. Food ads were categorized as permissible or not permissible for marketing to children using the World Health Organization's European Nutrient Profile Model. There were 1999 food ads in the movie theatre environment in Ontario and 43 food ads identified in the movie theatre environment in Virginia. On average, 8.6 (SD = 3.3) and 2.2 (SD = 0.9) food ads were displayed before children's movies in Ontario and Virginia, respectively. Most or all (97%-100%) food ads identified in Virginia and Ontario were considered not permissible for marketing to children. The results suggest that movie theatre environments should be considered for inclusion in statutory food marketing restrictions in order to protect children's health.

How much are we exposed to alcohol in electronic media? Development of the Alcoholic Beverage Identification Deep Learning Algorithm (ABIDLA)

BACKGROUND

Evidence demonstrates that seeing alcoholic beverages in electronic media increases alcohol initiation and frequent and excessive drinking, particularly among young people. To efficiently assess this exposure, the aim was to develop the Alcoholic Beverage Identification Deep Learning Algorithm (ABIDLA) to automatically identify beer, wine and champagne/sparkling wine from images.

METHODS

Using a specifically developed software, three coders annotated 57,186 images downloaded from Google. Supplemented by 10,000 images from ImageNet, images were split randomly into training data (70 %), validation data (10 %) and testing data (20 %). For retest reliability, a fourth coder re-annotated a random subset of 2004 images. Algorithms were trained using two state-of-the-art convolutional neural networks, Resnet (with different depths) and Densenet-121.

RESULTS

With a correct classification (accuracy) of 73.75 % when using six beverage categories (beer glass, beer bottle, beer can, wine, champagne, and other images), 84.09 % with three (beer, wine/champagne, others) and 85.22 % with two (beer/wine/champagne, others), Densenet-121 slightly outperformed all Resnet models. The highest accuracy was obtained for wine (78.91 %) followed by beer can (77.43 %) and beer cup (73.56 %). Interrater reliability was almost perfect between the coders and the expert (Kappa = .903) and substantial between Densenet-121 and the coders (Kappa = .681).

CONCLUSIONS

Free from any response or coding burden and with a relatively high accuracy, the ABIDLA offers the possibility to screen all kinds of electronic media for images of alcohol. Providing more comprehensive evidence on exposure to alcoholic beverages is important because exposure instigates alcohol initiation and frequent and excessive drinking.

Awareness of alcohol marketing, ownership of alcohol branded merchandise, and the association with alcohol consumption, higher-risk drinking, and drinking susceptibility in adolescents and young adults: a cross-sectional survey in the UK

OBJECTIVES

To explore awareness of alcohol marketing and ownership of alcohol branded merchandise in adolescents and young adults in the UK, what factors are associated with awareness and ownership, and what association awareness and ownership have with alcohol consumption, higher-risk drinking and susceptibility.

DESIGN

Online cross-sectional survey conducted during April-May 2017.

SETTING

The UK.

PARTICIPANTS

Adolescents and young adults aged 11-19 years in the UK (n=3399).

MAIN OUTCOME MEASURES

Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) (0-12) and indication of higher-risk consumption (>5 AUDIT-C) in current drinkers. Susceptibility to drink (yes/no) in never drinkers.

RESULTS

Eighty-two per cent of respondents were aware of at least one form of alcohol marketing in the past month and 17% owned branded merchandise. χ2 tests found that awareness of marketing and ownership of branded merchandise varied within drinking variables. For example, higher awareness of alcohol marketing was associated with being a current drinker (χ2=114.04, p<0.001), higher-risk drinking (χ2=85.84, p<0.001), and perceived parental (χ2=63.06, p<0.001) and peer approval of consumption (χ2=73.08, p<0.001). Among current drinkers, multivariate regressions (controlling for demographics and covariates) found that marketing awareness and owning branded merchandise was positively associated with AUDIT-C score and higher-risk consumption. For example, current drinkers reporting medium marketing awareness were twice as likely to be higher-risk drinkers as those reporting low awareness (adjusted OR (AOR)=2.18, 95% CI 1.39 to 3.42, p<0.001). Among never drinkers, respondents who owned branded merchandise were twice as likely to be susceptible to drinking as those who did not (AOR=1.98, 95% CI 1.20 to 3.24, p<0.01).

CONCLUSIONS

Young people, above and below the legal purchasing age, are aware of a range of alcohol marketing and almost one in five own alcohol branded merchandise. In current drinkers, alcohol marketing awareness was associated with increased consumption and greater likelihood of higher-risk consumption. In never drinkers, ownership of branded merchandise was associated with susceptibility.