Cigarette smoking remains the leading cause of preventable death in the United States [1], and over 37 million adults in the United States currently smoke cigarettes [2]. Although many adults who smoke make deliberate quit attempts each year, most attempts are unsuccessful [3–6] and utilization of evidence-based smoking treatments is low [7], thus more effective and scalable treatments are critically needed to enhance success in quitting. Quitting smoking is difficult in part because smoking is a highly automatic behavior [8–10] (e.g., lighting a cigarette without thinking about it), and automaticity of smoking is a central feature of nicotine dependence and a significant predictor of relapse [11]. Achieving and maintaining smoking abstinence requires, in part, being aware of smoking patterns and disrupting automatic behaviors to avoid smoking. Self-monitoring is an intervention strategy that is theorized to bring awareness to behavior and associated cues through event recording at the time that smoking occurs, thereby increasing perceived behavioral control and promoting behavior change, according to social-cognitive theories of behavior [12–15]. Evidence in support of this theory shows that self-monitoring is effective at reducing smoking [16–18]. However, problems with traditional self-initiated event monitoring (e.g., manually recording each smoked cigarette) include low adherence due to reasons such as fatigue, forgetting to record, and lack of real-time awareness of the behavior [19–21].

Wearable technology such as smartbands or smartwatches can provide an innovative solution to address cigarette smoking by providing automatic, continuous, and timely smoking assessment. These wearable devices are worn on the wrist and contain embedded geospatial sensors that use machine learning algorithms to identify the hand-to-mouth gestures associated with smoking a cigarette and notify the individual in real-time when smoking is detected, thereby bringing awareness to the otherwise automatic behavior of smoking. Several smartbands/watches have been initially validated to detect cigarette smoking in laboratory [22] and real-world settings [23–25], including using proprietary sensors [26,27] or off-the-shelf smartbands/watches [24,25,28]. Using smartbands to monitor and detect smoking may be advantageous due to their availability, ease of use, popularity, and potential for dissemination [29]. Preliminary research with the smartband used in the current study demonstrated high accuracy in detecting cigarette smoking (80–100% accuracy) with few false alarms (2–5%) indicating the smartband can reliably detect smoking and can differentiate smoking from other repeated hand gestures [28,30–32]. A pilot randomized controlled trial testing this smartband monitoring and notification system found a greater reduction in the number of cigarettes smoked per day after 30-days of wearing the study smartband compared to a wait-list control [31].

The current study built on this prior work by examining the feasibilty and preliminary efficacy of using a smartband to enhance quitting success among adults engaged in outpatient tobacco treatment. All participants received standard outpatient tobacco treatment as usual that included individual counseling and smoking cessation pharmacotherapy, and participants were randomized to receive either an active smartband for real-time smoking monitoring and notification (experimental group), or a control smartband (i.e., the same smartband but the smoking sensor was not active) for 8 weeks of treatment. The study aims were to: 1) evaluate the feasibility, acceptability, and helpfulness of the real-time smartband smoking intervention, and 2) assess the preliminary efficacy of the real-time intervention as an adjunct to standard tobacco treatment.

All study procedures were approved by the Yale University IRB. We recruited adults from two outpatient hospital clinics from 2020-2021 at Yale-New Haven Hospital in Connecticut, USA, that provide individual tobacco treatment to adults who smoke. New patients that were referred to the tobacco treatment clinics who arrived for an intake appointment to enroll in tobacco treatment were offered the opportunity to participate in the study via clinician referrals. Research staff contacted potential participants to determine initial eligibility, and those who were interested in participating in the research study to test the adjunct smartband treatment met with research staff to obtain written informed consent and complete questionnaires to confirm eligibility and enroll in the study (see Fig 1, consort diagram). All research study procedures were conducted remotely, and participants were compensated via pre-paid debit card in amounts of $20 for each research visit with an additional $20 study completion bonus for completing the week 8 visit and returning the smartband and charger via prepaid envelope.

This randomized clinical trial evaluated the use of a wearable smartband for supporting smoking treatment among adults engaged in tobacco treatment in an outpatient hospital setting. There is potential for using wearable technology to support health behavior change such as quitting smoking, given the current availability and projected growth in the wearable market [39–41]. Notably, 41% of U.S. adults currently own a wearable device such as a smartwatch or wearable health tracker [42], suggesting that intervention programs that use this technology could have wide population reach. Study findings provide important new information about using a wearable smartband to support quitting smoking and provide considerations for future research.

This study is the first to test the use of wearable technology for tracking smoking among an adult sample from an outpatient hospital setting. We enrolled an older adult sample on average, typical of an outpatient hospital setting, and this population of adults has known difficulties successfully quitting smoking, given high rates of medical comorbidities, long histories of established cigarette smoking, and higher than average nicotine dependence scores [38,43,44]. Yet the findings showed it is feasible and acceptable among an older population to use smartband technology and to adhere to it, as demonstrated by objective data from the smartband showing it was worn for most days (81.4%) out of treatment. These rates of adherence are consistent with other studies using this smartband among younger participants recruited nationally [45]. Additionally, participants reported high ratings for treatment satisfaction, helpfulness, and ease of use of the smartband, and the majority reported they would be willing to use it again and would recommend it to others. However, 50% of the sample reported it was difficult to remember to wear the smartband each day, identifying an opportunity for improving intervention adherence. We programmed automated text messages twice daily to remind participants to charge their smartband prior to their usual bedtime and to put on the smartband first thing in the morning at their usual wake time, and future studies may want to consider additional reminders to support consistent use of the wearable technology to optimize adherence and enhance the potential clinical benefit. Importantly, larger studies are needed to establish what thresholds for adherence and use of wearable technology for tracking smoking are related to clinical improvement, as adherence across wearable studies is mixed and not well defined [23].

Additionally, this study shows potential efficacy for reducing smoking by using a smartband for tracking smoking as an adjunct to standard tobacco treatment among patients in an outpatient hospital setting. We observed differences in smoking outcomes between groups, although this pilot study was not sufficiently powered to detect statistically significant differences. Biochemically confirmed 7-day PPA rates at the end of treatment were higher in the experimental group (n = 2/18 11%) compared to the control group (n = 1/20 5%), and those in the experimental group reported more percent days smoke-free and had larger reductions in cigarettes smoked per day during treatment. These quit rates are comparable to other pilot smartband studies, including a single-arm study testing a real-time mindfulness intervention delivered alongside smartband-detected smoking events [45] and a randomized two-arm study that compared standard tobacco treatment (counseling and nicotine replacement therapy) alone to standard treatment plus a smartwatch to detect smoking among a sample of adults with HIV who smoked cigarettes daily [46]. Additional fully-powered studies are needed to test the efficacy of smartband technology for tracking smoking and providing real-time intervention opportunities for the general population of adults who smoke.

These findings should be considered in light of study strengths and limitations. Notably, this study tested a highly novel technology in a new patient population (adults from an outpatient hospital setting) using a randomized controlled design that compared the active smartband to a sham smartband control, which adds to the scientific rigor. This is the first study that tested a sham smartband in the control group, which is a stringent control that accounts for demand characteristics associated with remembering to charge and wear the band and having smoking behavior monitored continuously with the band, just without receiving notifications. We did not assess participant beliefs about whether their smartband was active or sham, although it is possible that participants were able to ascertain their study group. Additionally, there might be other factors related to treatment effects that could be considered in future studies like use of counseling and medication adherence. Furthermore, the study smartband has not been tested to differentiate cigarette smoking from other smoking behaviors, such as cannabis or other tobacco product use, so we recruited adults who smoked cigarettes only and excluded for other tobacco product use or regular cannabis use. No other tobacco product use was noted during the study and only infrequent cannabis use was reported. Further work is warranted to test the ability of the gesture detection algorithm to differentiate cigarette smoking from other similar gestures, such as cannabis or other tobacco product use, to expand the use of this novel technology. Lastly, the study was conducted among a sample of patients from tobacco treatment programs in Connecticut, so findings may not generalize to other populations. However, the research study was conducted fully remotely, providing further support for the feasibility of using this technology broadly.

Future research might consider additional opportunities to expand on the use of this wearable technology to provide real-time support for quitting. For example, studies have used smartband data to deliver real-time interventions at the time of smoking (e.g., mindfulness exercises) and in advance of predicted smoking events based on individual smoking patterns obtained from smartband data [45]. Thus, this technology could be used to build on studies that have tested the use of smartband technology alongside or as an adjunct to tobacco cessation pharmacotherapy like nicotine replacement therapy [46,47] by testing ways to optimize intervention delivery with the smartband such as encouraging real-time medication use or practicing coping skills in the moment to avoid smoking. Furthermore, as the wearable market continues to grow, additional studies are needed to evaluate the dissemination potential of this technology by testing the smoking tracking and monitoring app system with adults who already own smartwatches, as this smoking monitoring app is compatible across iOS and Android systems and works with existing smartwatch technology. We found that some people reported difficulty remembering to wear the smartband daily and the smartband was worn on average for 8.5 hours (SD = 4.9) daily. Adherence may be improved by using wearable technology that someone already owns and is familiar with, as compared to remembering to wear a smartband just for the study. Although the smartband is not currently FDA-regulated as a medical device, future clinical trials could provide supporting evidence for this approval.

In conclusion, these results provide initial support for the feasibility, acceptability, and potential efficacy of using smartband technology to provide real-time smoking monitoring for adults in an outpatient hospital tobacco treatment program. Findings show promise that an older outpatient population can use this innovative technology with high rates of adherence and can benefit from real-time tracking of smoking behavior. As the wearable market continues to grow, there is wide potential to test the use of this technology to support health and behavior change.