Article Text
PDF
Abstract
Objective To determine whether awareness of emerging vaping-attributable health conditions influences vaping-related risk perceptions and behaviours among young people.
Design Respondents aged 15–24 years (n=3536) were drawn from a repeated cross-sectional online panel survey (222 participants/week) during an e-cigarette/vaping-associated lung injury (EVALI) outbreak in the USA (September 2019–January 2020). Logistic regression models tested for associations between EVALI awareness and perceived lung injury risk and product harm, stratified by e-cigarette/vape use and controlling for awareness of other e-cigarette/vaping news stories, demographic characteristics and outbreak week. Other models measured the association between perceived risk of lung injury and intentions to use (non-users) or intentions to quit (current users) e-cigarettes/vape products. Changes in national retail e-cigarette sales data were examined during national EVALI outbreak reporting.
Results EVALI awareness was associated with: perceived risk of lung injury (current users OR 1.59, p=0.004; non-users OR 2.11, p<0.001); belief that e-cigarettes/vapes contain dangerous chemicals (current users OR 1.47, p=0.017; non-users OR 1.88, p<0.001) and belief that e-cigarettes/vapes are harmful (current users OR 1.66, p=0.002; non-users OR 1.67, p<0.001). Perceived risk of lung injury from e-cigarette/vape use was associated with intentions to own e-cigarette/vape products (ever-users OR 0.25, p<0.001; never-users OR 0.61, p=0.004) and intentions to quit among current users (OR 2.02, p=0.002). Declines in e-cigarette sales were observed following news of the EVALI outbreak.
Conclusions News of vaping-attributable health conditions may prevent e-cigarette/vape use and encourage cessation among young people. Tobacco control campaigns should address uncertain health effects of e-cigarettes or vape products and align with risk communication by public health agencies during outbreaks.
- electronic nicotine delivery devices
- priority/special populations
- media
Data availability statement
Data are available on reasonable request. A data sharing agreement is required for use of all data. Investigators seeking access to data used in the study should make a written request to Truth Initiative authors and submit a detailed research plan including the purpose of the proposed research, required variables, duration of the analysis phase, IRB approval with FWA information, and investigator training in human subjects. Approved investigators may access datasets via an analytical Portal owned and administered by Truth Initiative.
Statistics from Altmetric.com
Background
Traditional tobacco products, such as cigarettes and smokeless tobacco, cause devastating health conditions and death among users.1 Tobacco-attributable diseases (eg, cancers, heart disease, stroke) have longer latency periods and appear years—often decades—after users initiate . Mass media campaigns that highlight these serious health effects have contributed to population-level declines in smoking prevalence by changing social norms and increasing public support for prevention and cessation of tobacco use.2–5 Among the most effective youth-targeted antismoking campaigns have been those that employ message frames including youth empowerment, industry manipulation, dangerous product designs (eg, chemicals) and the serious long-term health effects of smoking.6–9
In contrast, research examining the health consequences of new non-combustible tobacco products is limited. Communicating about novel outbreaks, such as life-threatening vaping-attributable health conditions, has not been studied by tobacco control researchers. It is unclear whether youth and young adults who become aware of respiratory conditions associated with vaping or e-cigarette use are deterred from using these products.
In August 2019, the Centers for Disease Control and Prevention (CDC), Food and Drug Administration (FDA), and state and local health departments launched a nationally coordinated outbreak investigation concerning hospital reports of patients across multiple states presenting with a novel, acute pulmonary condition.10 Most patients were young, with a median age of 24 years and 76% were under the age of 35.11 All reported use of e-cigarette or vaping devices, with the majority (82%) using products containing tetrahydrocannabinol (THC), the primary psychoactive component of cannabis (33% of whom exclusively used THC-containing products). Slightly more than half (57%) of patients used nicotine-containing e-cigarette or vape products (17% of whom exclusively used nicotine-containing products).11 The newly identified condition was, therefore, termed e-cigarette/vaping associated lung injury (EVALI). By 18 February 2020, 2807 hospitalised cases of the condition had been identified by CDC (figure 1), and 68 EVALI-attributable deaths had occurred.12
Cumulative confirmed and probable cases of e-cigarette/vaping-associated lung injury reported by CDC, September 2019–-February 2020. CDC, Centers for Disease Control and Prevention.
This national outbreak investigation was accompanied by formal risk communication efforts from health officials. Risk communication is the real-time dissemination of information by experts and advisors to the public during a crisis or emergency event, including guidance on protective behaviours and other recommended responses.13 Its immediacy amid rapidly changing circumstances distinguishes risk communication from other health communication efforts which use mass media campaigns or educational materials to promote social norms and behaviours . CDC and FDA held joint weekly press briefings and issued press releases on what was known and remained unknown about the lung condition, and steps being taken to address knowledge gaps.10 14 15 Over the course of the outbreak, CDC and FDA developed prevention recommendations, including refraining from the use of e-cigarettes or vaping devices, avoiding purchase of e-cigarette or vape products from sources who were not established retailers, and avoiding the use of vaping devices with THC.16 Risk communication efforts were covered by the news media and widely disseminated, often alongside other news about youth vaping behaviours and proposed policy responses.
At the time of the outbreak, health communication campaigns were already highlighting the potential for health consequences from vaping. Campaigns did not explicitly refer to the 2019 EVALI outbreak, but included information about the chemicals in these products and the lack of clear evidence related to their long-term health effects. The truth campaign aired its Tested on Humans ads using a metaphor of animal experimentation to question the ethics of allowing products with unknown health effects onto the consumer market. FDA’s The Real Cost campaign featured Why Is Vaping an Epidemic?, an ad depicting parasite-like creatures entering e-cigarette users’ bodies as a metaphor for nicotine’s systemic effects.
Examining youth and young adult responses to the 2019 EVALI outbreak can inform the development of effective communication strategies concerning new and emerging vape-related conditions. Multiple health behaviour theories posit that increases in health-related knowledge contributes to changes in attitudes and beliefs (eg, risk perceptions), which then lead to health behaviours .17 18 This study analysed survey responses among youth and young adults during the EVALI outbreak to determine whether (1) those who were aware of the outbreak as it was occurring were more likely to perceive greater personal risk of lung injury or e-cigarette product harm and (2) those with greater perceived risk of EVALI had weaker intentions to use, or stronger intentions to quit, e-cigarette/vape products in the future. We also examine changes in national retail e-cigarette sales during the period related to national reporting of the EVALI outbreak to lend additional evidence related to e-cigarette-related behaviour.
Methods
Data collection
News media scan
Tobacco Watcher, a real-time artificial intelligence service for tobacco control news media monitoring, was used to identify news reports of EVALI.19 Search filters were set according to product type (e-cigarettes), location (USA) and language (English). Tobacco Watcher results were compared with results from searches of an aggregated news site (Google News).20 Search terms included “vaping AND lung injury OR lung injuries,” “vaping AND hospitalization,” “vaping AND death(s),” “vaping associated lung injury” “EVALI,” “THC,” “THC AND lung injury,” and “vitamin E acetate.” Each week, two coders reviewed and categorized eligible news stories . Identified themes of the news stories were then developed into survey items for the weekly online panel survey.
Online panel survey
Data were drawn from a cross-sectional, continuous tracking survey of participants from the national Dynata (formerly Research Now) online panel (https://www.dynata.com). Approximately 222 participants (aged 15–24 years) are surveyed per week. Participants who completed a survey between 25 September 2019 and 14 January 2020 were included in the analytical sample (n=3536). Data were weighted to be nationally representative using US Census demographic benchmarks.
Retail e-cigarette sales
Aggregated 4-week e-cigarette sales at the national level were used as an indicator of consumer behaviour during and following the EVALI outbreak. Nielsen retail scanner data estimates e-cigarette product sales (eg, e-cigarette devices, cartridges, e-liquids) from: independent, chain and gas station convenience stores; food, drug and mass merchandisers; discount and dollar stores and military commissaries. The dataset does not include online or small independent business sales, including vape shops.
Measures
Predictors
Awareness of EVALI news
E-cigarette-related or vape-related news topics were drawn from weekly scans of news stories and summarised in a brief sentence for inclusion on the survey. A total of 19 news story statements were included in the survey, with the order randomised for each respondent. Table 1 presents the news story statements used in the analysis. Statements were removed from the survey instrument as they became outdated, while others were retained over time.
E-cigarette/vape news story statements during the EVALI outbreak period
Awareness of news story statements were measured using the following prompt: ‘Which, if any, of these news items have you seen in the media in the past week?’ For each news story statement, respondents indicated either ‘I HAVE seen this in the media in the past week’ or ‘I have NOT seen this in the media in the past week’.
An exploratory factor analysis of survey data identified correlations between news story statements, then assessed for face validity by the research team. Statements were categorised according to message frames: EVALI Identification (unknown or emerging aspects of the lung injury outbreak); EVALI Cause (hypothesised or identified causal mechanisms of lung injuries among e-cigarette/vape users); and/or Product Bans (policies to restrict access or remove e-cigarettes from the consumer market).
Three dichotomous measures of awareness of each news category (EVALI Identification, EVALI Cause and Product Bans) are defined according to whether the respondent reported seeing any news statement that fell under the defined news category (table 1). For example, if the respondent had seen the news statement, ‘Some states and local governments are working to ban flavoured vape/e-cigarette products,’ respondents were coded as having awareness of the Product Bans category.
Outcomes
Perceived risk
Respondents were asked to indicate their agreement with a time-bound scenario about their risk of lung injury following e-cigarette/vape use (‘If I were to vape/use e-cigarettes now, the following could happen to me within the next year: [My lungs could be seriously injured]’). Agreement was measured on a 5-point Likert-type scale and dichotomised for analysis (0=strongly disagree, disagree, or neither agree nor disagree; 1=strongly agree or agree).
Perceived product harm
Respondents were asked to indicate their agreement with two statements about whether e-cigarette/vape products: ‘contain dangerous chemicals’ and ‘are harmful to your health’. Agreement was measured on a 5-point Likert-type scale and dichotomised for analysis.
Behavioural intentions
Respondents who were not current (ie, past 30-day) users were asked, ‘Do you think you will own a vape, JUUL, or other e-cigarette in the next year?’. Response options were presented on a 4-point Likert-type scale and dichotomised for analysis (0=definitely not or probably not, 1=definitely yes or probably yes).
A measure of quit intentions among current e-cigarette or vape users was added to the survey on 23 October 2019. The item asked respondents, ‘Are you seriously thinking about quitting e-cigarettes/vapes?’. Responses options reflected the progression from no intention to quit to current quit attempt: ‘No, I am not thinking of quitting’, ‘Yes, but not within the next year’, ‘Yes, within the next year’, ‘Yes, within the next 6 months’, ‘Yes, within the next 30 days, and ‘I’ve already quit’ and dichotomised for analysis (0 = ‘No, I am not thinking of quitting’; 1=any ‘Yes,…’ response).
Covariates
Current use
Self-reported past 30-day use of e-cigarette/vape products is used as a dichotomous measure of current use. Respondents were asked, ‘During the past 30 days, on how many days did you use each of the following products?’: JUUL; e-cigarettes/vapes (not including JUUL). Respondents who reported between 1 and 30 days of using either or both products were coded as current users.
Ever use
Respondents were asked, “Have you ever used or tried any of the following tobacco products, even one or two times?” : JUUL; e-cigarettes or vapes (not including Juul). Respondents who answered yes to either product were coded as ever users.
Demographic characteristics
Demographic covariates include age (0=15–17 years or 1=18–24 years), race/ethnicity (0=white or other non-Hispanic; 1=Black non-Hispanic; or 3=Hispanic), and gender (0=male, 1=female). A measure of perceived financial circumstances is used as a proxy for household income, measured by asking ‘How would you describe your/your family’s overall financial situation?’. Response options were: 0=‘don’t meet basic expenses,’ 1=‘just meet basic expenses with nothing left over,’ 2=‘“meet basic needs with a little left over,’ or 3=’live comfortably.’21
Outbreak week
The week of each respondent’s survey completion is used to control for unknown confounders of news story awareness, attitudes and changes in the outbreak over time (weeks of 25 September–1 October 2019 to 8 January–14 January 2020).
Retail e-cigarette sales
Per capita sales dollars were constructed using monthly population figures from the US Census Bureau for the civilian population aged 18 years and older and adjusted for inflation to March 2020 dollars. An interrupted time series with per capita Nielsen sales data (July 2017—March 2020) was used to model the impact of CDC’s announcement of the EVALI outbreak investigation (6 September 2019) on the sale of e-cigarette products. We used multiple treatment periods to control for events that could also affect the sales of e-cigarettes. The first treatment period is defined as the time span between Rite Aid’s announcement that the company would cease e-cigarette sales (11 April 2019) through the release of CDC’s announcement (6 September 2019). The second treatment period is defined as the time span between CDC’s announcement (6 September 2019) through JUUL’s announcement that it would cease the sale of mint JUUL pods (7 November 2019).22
Analysis
Survey data
Separate multivariable logistic regression models measured the association between awareness of news media category (EVALI Identification, EVALI Cause and Product Bans) and the following outcomes: (1) perceived risk of lung injury; (2) belief that e-cigarettes/vapes contain dangerous chemicals and (3) belief that e-cigarettes/vapes are harmful to health. Additional models were constructed to determine the association of perceived risk of lung injury and intention to own an e-cigarette or vape product (among ever-users or never-users) or intention to quit e-cigarette or vape use (among current users). Each model controlled for age category, race/ethnicity, gender and perceived financial circumstances. A fixed effect for outbreak week controlled for constant within-week heterogeneity. Models examining risk and harm perceptions were stratified by current or non-current e-cigarette or vape use, and the model examining intention to own an e-cigarette or vape was stratified by ever or never use.
Nielsen data
Analysis was conducted in Stata version 15.1 using the program ITSA. To control for autocorrelation, Newey-West robust standard errors were used.
Results
A total of 3536 respondents completed the survey. One-quarter (25.6%) were aged 15–17 years and three-quarters (74.4%) were aged 18–24 years (table 2). Slightly less than three-quarters (71.5%) of respondents were white non-Hispanic or some other race, with the remainder reporting Black non-Hispanic race (12.3%) or Hispanic ethnicity (16.3%). The sample was evenly distributed according to binary gender (51.2% male). Slightly less than half of the sample reported lower perceived financial circumstances, either not meeting basic expenses (11.6%) or just meeting basic expenses with nothing left over (32.5%).
Respondent characteristics (n=3536)
The proportion of respondents who reported ever using or trying an e-cigarette/vape product (40.9%) or being current users (29.1%) were similar to 2019 national estimates.23 One-quarter of current (past 30-day) users (26.0%) expressed no intention to quit (table 2).
Perceptions of lung injury risk and e-cigarette/vape product harm
Awareness of news story statements that described the emergence of EVALI (‘EVALI Identification’) was associated with perceived risk of lung injury among both current users (OR: 1.59, p=0.004) and non-users (OR: 2.11, p<0.001) of e-cigarette or vape products (table 3). For current users, this association was sustained when controlling for awareness of information about EVALI causal mechanisms (ie, vitamin E acetate in some THC-containing vape products) as they were discovered (‘EVALI Cause’), as well as news stories about policy responses (‘Product Bans’), both of which had non-significant effects on perceptions of lung injury risk. For non-users, awareness of information about policy responses (‘Product Bans’) had a significant positive association with perceived lung injury risk (OR: 1.29, p=0.033).
Logistic regression models of the association between EVALI awareness and perceived lung injury risk and product harm perceptions among youth and young adults
Similarly, awareness of the EVALI Identification message frame was predictive of the belief that e-cigarettes/vapes contain dangerous chemicals among current users (OR: 1.47, p=0.017) and non-users (OR: 1.88, p<0.001), and the belief that e-cigarettes/vapes are harmful to health by current users (OR: 1.66, p=0.002) and non-users (OR: 1.67, p<0.001) (table 3). Across these models, age was a significant covariate only among non-users, with young adults having significantly lower odds of expressing EVALI risk and product harm perceptions compared with youth. The odds of perceiving lung injury risk and harmfulness of e-cigarette/vape products were significantly higher among Black or Hispanic current users compared with white non-Hispanic current users and female respondents compared with males, but lower among Black or Hispanic non-users compared with white non-Hispanic non-users (table 3).
Behavioural intentions
Respondents with perceived risk of serious lung injury from e-cigarette or vape use had lower odds of expressing intentions to own an e-cigarette or vape product in the future, among both ever-users (OR: 0.25, p<0.001) and never-users (OR: 0.61, p=0.004) of e-cigarette or vape products (table 4). Among current users, those who perceived risk of lung injury had twice the odds of expressing quit intentions (OR: 2.02, p=0.002) compared with those with lower risk perceptions.
Logistic regression models of the association between perceived risk of lung injury and e-cigarette or vape-related behavioural intentions among youth and young adults
Holding perceived risk of lung injury constant, the odds of intending to own an e-cigarette or vape product were significantly higher among Black non-Hispanic (OR: 1.85, p=0.001) and Hispanic (OR: 1.56, p=0.018) respondents who had never used an e-cigarette or vape product compared with white non-Hispanic never-users. In the model testing the relationship between perceived risk of lung injury and quit intentions, Black non-Hispanic current users had more than twice the odds of intending to quit (OR: 2.41, p=0.046) compared with white non-Hispanic current users.
E-cigarette sales
The starting level of real per capita e-cigarette sales was estimated at 30 cents. Per capita sales increase every month prior to the Rite Aid announcement by 6 cents (p<0.001). In the first month after the Rite Aid announcement, there is a non-significant decrease in per capita sales of 13 cents (p=0.082), followed by statistically significant monthly trend increase of 9 cents per capita (p<0.001) until the CDC announcement. CDC’s 6 September 2019 outbreak investigation announcement is associated with an initial statistically significant decrease of 12 cents per capita (p<0.001), followed by a statistically significant monthly trend decrease of 3 cents per capita (p<0.001) until JUUL’s announcement. Lastly, JUUL’s announcement is associated with an initial statistically significant increase of 40 cents per capita (p<0.001), followed by a statistically significant monthly trend decrease of 7 cents per capita (p<0.001) (figure 2).
Multiple treatment interrupted time series evaluation of per capita sales of e-cigarettes following the CDC announcement of the 2019 EVALI outbreak investigation, compared with Rite Aid and JUUL announcements of e-cigarette product removals. CDC, Centers for Disease Control; EVALI, e-cigarette/vaping-associated lung injury.
Discussion
Findings from this study demonstrate that awareness of EVALI was associated with perceived risk of e-cigarette or vape use among both current users and non-users. This association persisted even as respondents learnt of the likely causative agent of EVALI, vitamin E acetate as a liquid diluent in THC-containing vape products.24–26 THC vape products—loosely regulated and illegal in most US states—are not among the most commonly used e-cigarette or vape products among young people.27 This finding suggests that many young people were not distinguishing between e-cigarette or vape ingredients or devices when assessing EVALI risks. Knowing that any e-cigarettes or vape products were implicated in lung injuries appears to have been sufficient to increase youth and young adults’ harm perceptions of these products.
Data suggest that announcements about EVALI by government health agencies were associated with a sustained decrease in e-cigarette sales. This level of decline was not observed after a major pharmacy chain announced it would no longer sell e-cigarettes, nor after a major e-cigarette company announced it would remove flavoured e-cigarettes from retailers. The decline in sales on the launch of the CDC outbreak investigation—a period defined by high uncertainty and concern about the cause of lung injuries in e-cigarette and vape users—is consistent with the survey results which demonstrate that fears about lung injuries predict lesser intentions to own e-cigarettes and greater intentions to quit.
While the EVALI outbreak appeared to have an effect in significantly reducing sales, we are unable to determine whether this reduction would have been sustained without subsequent changes in the e-cigarette market. These changes included JUUL’s voluntary removal of advertising and flavours with youth appeal in anticipation of more stringent federal regulation on e-cigarettes. EVALI was a precipitating event that increased public scrutiny and pressure on e-cigarette manufacturers and policymakers to address the dramatic increase in youth e-cigarette use.
Given the lack of evidence related to long-term health consequences for new and emerging tobacco products, it remains possible that lung injury outbreaks could become a recurring issue.28–31 Risk communication efforts can leverage message frames used by campaigns that effectively and efficiently communicate health risks to youth and young adults. Similarly, health communication campaigns can incorporate the uncertainty surrounding e-cigarettes into prevention and cessation messages to reinforce risk communication by health officials. The results of this study suggest that the high degree of uncertainty surrounding the potential harms of e-cigarette or vape use may be especially relevant motivators of behaviour change for youth and young adults during outbreaks of vaping-related pulmonary conditions.
Limitations
Although this study has many strengths, it is not without limitations. Nielsen e-cigarette sales are not limited to sales among young people. However, youth and young adults report the highest prevalence of e-cigarette use across all age groups.32 33 Nielsen data do not account for all potential points of sale, omitting online sales and independent retailers including vape shops. It is unknown whether sales systematically differed from these sources compared with other retail types during the EVALI outbreak.
The analytical sample of the cross-sectional youth and young adult survey was restricted to a date range that preceded the arrival of COVID-19 in the USA. However, the final 4-week data point of the Nielsen sales data encompasses approximately 1 week by which most states had issued emergency declarations about COVID-19, and the beginning of stay-at-home orders (ie, 14 March –21 March 2020). While e-cigarette sales may be affected by the COVID-19 pandemic, approximately 90% of Nielsen data in this study reflect e-cigarette sales from convenience stores, which are deemed essential businesses during the COVID-19 pandemic and have remained open to consumers. Models run without the final data point demonstrate a similar effect, but its inclusion improves the precision of the model and thus was retained.
Survey results are cross-sectional, and therefore, unable to establish a causal direction between self-reported exposure to news reports and increases in risk perceptions and associated behavioural intentions. This study is strengthened by the measurement of EVALI awareness using rapid weekly tracking data and continuously updated contemporary news media reports airing each week.
Conclusions
EVALI awareness was associated with perceived risk of lung injury and beliefs about the harmfulness of e-cigarette and vaping products. Young people with perceptions of lung injury risk and product harm had significantly lower odds of intending to purchase an e-cigarette or vape product in the future, and among current users, greater odds of intending to quit. E-cigarette sales declined significantly during the period when national health officials were disseminating risk information about the potential role of e-cigarette and vaping products for lung injuries. Tobacco control campaigns targeting youth and young adults should incorporate messages concerning the uncertain health effects of e-cigarette and vape products and be prepared to align messages with public health agencies during outbreaks.
What this paper adds
This paper used rapid weekly tracking data to measure young people’s awareness of the e-cigarette/vaping associated lung injury (EVALI) outbreak. The study found associations between EVALI awareness, perceived risk and beliefs about product harm, and intentions to own or quit e-cigarettes.
The study includes an analysis of e-cigarette sales which indicates a decline in sales following the announcement of a national EVALI outbreak investigation.
Understanding youth and young adult responses to the 2019 EVALI outbreak can inform the development of effective risk communication strategies concerning new and emerging vape-related conditions.
Data availability statement
Data are available on reasonable request. A data sharing agreement is required for use of all data. Investigators seeking access to data used in the study should make a written request to Truth Initiative authors and submit a detailed research plan including the purpose of the proposed research, required variables, duration of the analysis phase, IRB approval with FWA information, and investigator training in human subjects. Approved investigators may access datasets via an analytical Portal owned and administered by Truth Initiative.
Ethics statements
Patient consent for publication
Ethics approval
The study received approval from the Advarra Institutional Review Board.
References
Footnotes
Twitter @MeganCDiaz, @ehairphd
Contributors JMK conceptualised the study. JMK designed and led the analysis of the survey data and MCD designed and led the analysis of the e-cigarette sales data. MS prepared e-cigarette sales data for analysis. ECH and DV provided oversight of data collection and guidance on analysis and interpretation. JMK drafted the manuscript with written contributions by MCD and DV and edits by MS and ECH.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Author note The conclusions drawn from the Nielsen data are those of the researcher(s) and do not reflect the views of Nielsen. Nielsen is not responsible for, had no role in, and was not involved in analyzing and preparing the results reported.