Reaching Intermittent Tobacco Users With Technology: New Evidence

Waterpipe (hookah) smokers, unlike cigarette smokers, are typically intermittent users who are rarely motivated to quit, making them difficult to reach with interventions. Mays et al. (p. 1686) conducted a randomized controlled trial of a six-week tailored text message intervention to reduce waterpipe smoking in young adults. Intent-to-treat results revealed that 43% of participants in the tailored text condition self-reported waterpipe cessation at six months versus 35% in an untailored text condition and 28% in an assessment-only control condition. Such a simple, inexpensive intervention that can easily be disseminated has high potential for impact on this type of tobacco use. Although many existing evidenced-based tobacco-cessation services are delivered in the health care system, interventions that operate outside the health care system are needed, particularly for young adults who may not regularly interact with health care providers and who may not even consider themselves smokers.

One challenge for implementing this type of intervention in the real world will be in determining how to engage waterpipe smokers who are not usually actively seeking to quit. Participants in the study by Mays et al. responded to advertisements for a study about “waterpipe tobacco beliefs and behavior” and received $100 for completing assessments. In the absence of financial incentives, people who are not ready to change may be reached by engaging them on a topic that resonates with their interests and values. For example, we reached mothers of adolescent daughters to reduce their permissiveness to allow their daughters to use tanning beds by creating a Facebook group–delivered campaign on mother–daughter communication about adolescent health, a topic of high interest to mothers.¹ The campaign involved twice daily posts for a year but a fraction (15%) of those posts (approximately two per week) were on the topic of tanning beds. Findings revealed that the campaign was successful in reducing mothers’ permissiveness in allowing their daughters to use tanning beds.² Text-based interventions could use this “embedded messaging” approach so that the text campaign could be on a highly engaging topic for young adults and include a small proportion of messaging on waterpipe smoking.

The bulk of text messages in both text conditions in Mays et al. provided education about the health harms and addictiveness of waterpipe tobacco use. Such education helps correct commonly held misconceptions about the risks of this type of tobacco use. However, future studies could incorporate other strategies known to affect long-term change in tobacco use, such as setting a quit day, refusal skills, coping strategies for cravings, and access to tobacco-cessation medication when appropriate. A recent clinical trial for e-cigarette cessation in young adults demonstrated that a text intervention including these strategies was effective at promoting vaping cessation.³

Even though much intervention content was educational, the tailored text messages of Mays et al. were based on baseline characteristics as well as brief answers to questions, primarily regarding the participant’s knowledge of waterpipe health risks and addictiveness. Interestingly, the tailored text condition did not outperform a nontailored text condition, which is in line with the extant literature indicating that tailoring has not consistently increased effects of text-messaging interventions.⁴ Recently, interest has grown in using just-in-time adaptive intervention approaches to improve the effects of tailored interventions by tailoring content to coincide with moments of opportunity, receptivity, or vulnerability.^5,6 If momentary tailoring variables for waterpipe use could be identified and monitored with mobile technology (e.g., presence at a party, urges to use), messages could be timed to engage momentary targets (e.g., peer pressure, cravings) on a just-in-time basis. Building on the notion of context sensitivity, future research should explore whether a generic, one-size-fits-all decision rule is optimal for text-messaging interventions to reduce waterpipe tobacco use. Different people may benefit from different messages at different times; that is, the optimal dosing of messages may be person specific, and person-specific decision rules may be required to optimize intervention effects.⁷

Although the text intervention decreased waterpipe smoking, the majority of the sample in Mays et al. also used at least one other tobacco product at baseline in addition to waterpipe tobacco. Specifically, 29% were current cigarette smokers and 68% had used a tobacco product other than waterpipe or cigarettes in the past 30 days. This is a strength of the study, as it likely accurately reflects the frequency of other tobacco product use among waterpipe smokers in the United States.⁸ However, cessation outcomes reflect only waterpipe smoking, and the use of other tobacco products is neither incorporated into the definition of cessation nor reported at follow-ups. The possibility exists that those who reduced waterpipe use concurrently increased use of other tobacco products, making the effect of the text interventions on overall tobacco risk profile unclear. The possibility of compensatory tobacco use may have been increased by the fact that the content of several of the intervention text messages sought to raise the perceived harm of waterpipes relative to other tobacco products (e.g., one text to participants read, “The large amount of smoke from hookah delivers more cancer-causing chemicals than cigarettes”). Future studies should examine how the interventions affected the perceived risk of waterpipe versus that of other tobacco products and if such changes in relative risk drive compensatory tobacco use.

Because of how waterpipe cessation was defined, caution is warranted when comparing the cessation rates and effect sizes reported by Mays et al. to other tobacco-cessation interventions. An intent-to-treat cessation rate of 43% is reported at six-month follow-up for the tailored text intervention, representing an odds ratio of 1.9 relative to assessment-only control. By itself, this seems remarkable for a low-cost tobacco-cessation intervention. However, the majority of other recent clinical trials targeting combustible tobacco define cessation as abstinence from the target combustible tobacco product, as well as other (i.e., not specifically targeted) combustible tobacco products.⁹ Recent guidelines regarding clinical trials of combustible cigarette cessation from the Society for Research on Nicotine and Tobacco recommend (1) reporting the use of all tobacco products at outcome assessments, and (2) defining cessation of combustible cigarettes as abstinence from combustible cigarettes and all other combustible tobacco products.⁹ Although these guidelines are about trials of combustible cigarettes, much of the rationale for these recommendations applies to waterpipe use as well. Defining cessation as abstinence from all combustible tobacco products would also allow biological verification of short-term abstinence, which is the current gold standard in tobacco cessation.¹⁰

Overall, we are cautiously enthusiastic about these new findings. This intervention addresses an important unmet need for waterpipe cessation and could be a valuable component of a multicomponent tobacco-cessation program. Given the proliferation of ways to smoke tobacco and consume nicotine, further work is needed to determine whether interventions targeting one form of tobacco use have a net harm reduction effect considering all other forms as well.