This study showed that exposure to a message communicating reduced exposure to harmful chemicals for AWS who switch completely to JUUL had the intended effect of increasing openness to try and use JUUL among AWS. At the same time, the data did not support the concern that such communication would increase openness to try or use JUUL among nonusers of tobacco, i.e., exposure to the message significantly reduced openness to trial among nonusers. Thus, communication of reduced exposures is likely to increase the likelihood that AWS use JUUL. This likely would yield a public health benefit, as multiple studies show that AWS using JUUL have a high probability of switching completely away from smoking: A large observational study found that, one year after adopting JUUL, 51.2% of baseline established smokers had not smoked at all in the previous 30 days [35]. And, among those still smoking, 60% had reduced their cigarette consumption by at least 50% [36], which is also associated with substantial reductions in exposure to toxicants [10]. A third study, in African-American and Hispanic AWS, showed complete switch rates of 28.1% after 6 weeks [37].
Mediational analyses confirmed the precepts of behavioral theories suggesting that information about health effects would not affect intention to change behavior directly, but, rather, through its effects on risk perceptions [16–19]. In this study, the effect of message exposure on AWS’ OTU was completely mediated by the message’s effect on risk perceptions. This highlights the important role of risk perceptions on influencing behavior and highlights the need to correct the misperceptions that are increasingly prevalent among AWS [20]. The Center for Tobacco Products at the US Food and Drug Administration has recently announced public communication initiatives to correct such misperceptions [38]. A limitation of the current study is that risk perceptions and openness were assessed at the same timepoint; future mediation analyses assessing openness at a latter timepoint could confirm this casual pathway.
The tested message addressed exposures to both AWS who switched to JUUL, as well as to bystanders, affecting risk perception for both [24]. Perceived harm to others also independently affected AWS’ OTU, consistent with literature indicating that AWS are concerned about the effects of their smoking on others, particularly family members [39, 40]. This aspect of harm-reduction from switching from smoking to ENDS may warrant further attention.
This study also evaluated how perception of message credibility impacted behavioral intentions. As reported in [24], in spite of having been exposed to the reduced-exposure message, Message Rejectors perceived JUUL as carrying greater harm to health than Control participants who had not seen the message. The present analyses showed that Message Rejectors were also less likely to be OTU than Control participants. However, mediational analyses showed that their risk perceptions did not explain Message Rejectors’ lower likelihood of OTU. This suggests that Message Rejectors were not just doubting the facts stated in the message, but rather that other factors were negatively influencing their OTU. Perhaps pre-existing attitudes and beliefs about ENDS drive Rejectors’ risk perceptions and lack of openness to using JUUL. As JUUL in particular has been the subject of much criticism in the media, attitudes about JUUL – the sponsor of the advertisement in which the message was embedded – may also have triggered negative responses. Unfortunately, the study did not evaluate attitudes and beliefs prior to message exposure, so this explanation cannot be confirmed in the data.
Numerically, the effects of the message on OTT and OTU were small. The impact may have been limited because Openness is such a low-threshold criterion for assessing behavioral intentions: a respondent who indicated “Strongly disagree” on all but one item was considered Open. The result was that 89% of AWS were considered OTT and 86% OTU even in the Control condition. These figures do not reflect the proportion that are likely to actually try or use JUUL, but the relative difference between the Message and Control conditions is expected to capture the relative increase in trial and use.
Another factor that may have limited the impact of the message was that participants were only required to view the message once during the study, whereas implementation would involve broader dissemination and presumably multiple exposures, possibly supported by more authoritative sources of health information. In a sense, the study is more a proof-of-concept than an estimate of the practical effect in actual implementation. In any case, even small increases in adoption of ENDS by AWS, when played out over time, can have substantial public health benefit, as demonstrated in population modeling based on data such as these [15].
The study findings help assuage concerns that a reduced-exposure message could promote behavior that would adversely affect public health. Exposure to the message significantly decreased nonusers’ openness to trying JUUL. Also, seeing the message did not interfere with AWS’ intention to quit smoking. And, among the few tobacco nonusers who indicated an openness to use JUUL, exposure to the message actually decreased susceptibility to smoking, addressing concerns about the message promoting progression to smoking. These findings mitigate the likelihood of adverse effects that would undercut the expected benefit of helping AWS to reduce their exposures to toxicants.
Several study limitations should be considered. The analyses focused on openness to use as a dichotomy, parallel to analyses of susceptibility to smoking [28], but this limited variance and yielded high prevalence of openness among AWS even in the Control group. The rates of OTU should not be interpreted as actual probability of use, though relative differences in Openness due to message exposure likely apply to actual use. The sample was not necessarily representative of the reference populations, but was diverse with respect to demographics and tobacco use history. The single exposure to the message may underestimate the impact that might be seen with repeated exposure. As advertisements have low credibility [41], and tobacco companies have low credibility [42], the fact that the stimulus was an advertisement from the manufacturer of JUUL likely also limited its credibility. Indeed, the data showed that respondents who rejected the message as not credible were less likely to be OTU. AWS’ behavior is likely to be more affected by messaging from credible health authorities, which is encouraging for the potential of FDA’s planned campaign to address misperceptions [38].
Strengths of the study include a large and diverse sample with representation of adults with varying tobacco use histories. The randomized experimental design provided for clear causal inferences about the effect of the message per se. Mediational analyses added to these causal inferences by considering the process by which the message and message credibility affected the outcomes considered, consistent with behavioral theory [16–19]. The message itself was realistic, in that it is well-supported by science, and mirrors a message already authorized by the FDA for a different reduced-exposure product [43]. Finally, the analyses addressed not only potential positive effects of message communication, but also considered potential adverse effects, showing no increased likelihood of trial or use among nonusers of tobacco, no evidence of diverting AWS from quitting, and no evidence of gateway effects.
Overall, the findings suggest that communication of reduced exposure to harmful chemicals for AWS who switch from smoking to JUUL is likely to produce a public health benefit, by encouraging AWS to switch to a lower-exposure product, without promoting use by nonusers of tobacco.