In the absence of empirical evidence to inform optimal mHealth program design, it is critical that clinical investigators and software developers consider the needs and preferences of smokers. This alone will not ensure treatment effectiveness, but it is an important consideration for treatment engagement, and engagement is necessary for efficacy. Surprisingly little published work has engaged smokers in the design process [28-30], yet important lessons can be learned from those who have. For example, Bock and colleagues [28] engaged smokers in the design of a text-based cessation program and learned smokers wanted a broader program that included social networking and control over the timing and delivery of text messages. Similarly, Ploderer and colleagues [30] field-tested a mobile app and discovered unanticipated strategies smokers use to cope with cravings and leverage peer support. Similarly, Paay and colleagues [29] engaged smokers in focus groups and design workshops and gained insight for tailoring smoking cessation apps to accommodate individual differences among smokers.

Other studies have not engaged smokers in the design process, but gained important information about smokers’ needs and preferences from observing their adoption and use of mHealth tools [31]. For instance, Heffner and colleagues [32] observed smokers’ use of a mobile quit-smoking app over 8 weeks and discovered that the most popular features were not necessarily associated with quitting. This underscores an important lesson—user preference does not guarantee treatment outcome. This is why it is important to ground design considerations in solid treatment theory and empirical evidence, but user preferences can offer important insights about how best to do this.

Solicitation of user preferences may be most effective when guided by existing knowledge gaps in the literature on messaging, social support, and incentivizing use. For example, it is known that text message reminders may be effective cessation aids [9,10] and may be useful for delivering tailored motivational messages [33,34], quit advice based on expert recommendations [10,35], and support on demand. However, text reminders about quitting smoking may also prompt people to smoke [29], so some smokers may not be receptive to this messaging. Some smokers prefer to control the timing of automated messages [28], whereas others prefer messages sent at unexpected times [35]. Smokers’ relative preferences for push messages received automatically versus pull messages delivered upon request has been largely unexplored [31], but is an important issue that warrants further exploration.

Similarly, social support is an important feature in smoking cessation programs [7,36]. Several technology-based strategies can be found in the literature, including mobile quit buddies [9], role model videos delivered via text message links [37], or public pledges posted on social media [15,38]. Research highlights unresolved tensions between the diffusion of quit-smoking information and experiences through online social networks [39-42] versus smokers’ desire to maintain anonymity [29,43]. These gaps also warrant further investigation to better understand how smokers wish to use mHealth technologies for social support during the quit process.

Finally, insight into smokers’ preferences for strategies that can make mHealth cessation aids rewarding and engaging can fill gaps in knowledge about how to best incentivize program use. Gamification is the use of game design elements in nongame contexts [44], such as health [45]. Some mHealth interventions pair game features with desired behaviors [46], such as encouraging physical activity by growing a digital garden as fitness goals are attained [47] or scoring points on familiar games (eg, hangman, Sudoku) for accurately estimating nutritional value of foods [48]. Gamification is well suited for motivational aspects of behavior change [49-51]. As a smoking cessation aid, game mechanics (eg, points, badges, leaderboards) have been used to encourage the use of educational “quit guides” [38]. Such game-based incentives can be integrated and viewable to others via social media [24,38]. However, others have advocated offering smokers more personally meaningful rewards, such as money [29]. Although prior work has elicited opinions of experts [52], smokers’ relative preferences for the use gamification features have been largely unexplored to date despite enthusiasm for their use by developers.

In summary, user-centered studies that engage smokers have proven valuable to uncovering important design considerations from the user’s perspective [28-30]. Given gaps in our understanding of smokers’ preferences for the range of possible mHealth support, further research is warranted to characterize their design priorities. In this study, we employed user-centered mixed methods to characterize the design preferences of smokers. Given the considerable opportunities to advance quit-smoking tools with mHealth innovations, such engagement of smokers in the design process is critical for identifying the most promising design features for future mHealth cessation tools.

We recruited smokers in the Seattle area via online ads (eg, Craigslist), posted flyers, and invitation letters mailed to likely smokers who were members of Group Health Cooperative (a nonprofit health care system in Washington State). Interested smokers were invited to contact study staff to learn more and be screened for eligibility. Individuals were eligible if they were at least 18 years old, a current smoker interested in quitting, could read and write in English, and owned a mobile phone or tablet computer which they used to access the Internet. We purposively recruited by age for three groups younger than 40 years of age and three groups 40 years of age and older.

Each focus group lasted approximately 2 hours and was moderated by the lead author (AH) with support and note taking by a coauthor (JB). We began each group with a paper survey to collect participant characteristics, including demographics (ie, race, ethnicity, education, employment status, income), smoking status (ie, cigarettes smoked in past 30 days, smoking frequency, plans to quit), and experience with technology (ie, some experience, intermediate, very experienced, expert). Then, participants engaged in a two-part group discussion. The first part focused on blue-sky brainstorming about the ideal design of future mHealth quit-smoking tools (part I). In the second part (part II), participants were asked to react to contrasting design options for mHealth messaging, social support, and incentivizing program use. At the end of the session, after participants had considered a range of design options for quit-smoking tools, they completed an exit survey rating their perceived importance of various mHealth features. Group discussion was audio recorded and transcribed for analysis, and field notes were written after each group meeting. Attendees received US $50 for their participation.

We conducted a mixed-methods analysis of qualitative focus group transcripts and quantitative surveys. Following each focus group session, the two facilitators (AH, JB) met to summarize key themes that emerged. After completion of all six groups, the facilitators (who are both trained in qualitative interviewing and coding methods) reviewed field notes and applied thematic analysis [54] to transcripts to identify qualitative themes regarding participants’ blue-sky design ideas and preferences for messaging, social support, and gamification features of mHealth quit-smoking tools. Each facilitator summarized themes in transcripts, which were then reviewed by the second facilitator for accuracy. Any discrepancies were resolved during meetings between the coders. An initial synthesis of themes depicting the data were reviewed with study team members. This coding scheme was applied to individual transcripts by AH capturing representative quotes for the themes. Themes were compared across the six focus groups and between age groups.

We summarized participant surveys with descriptive statistics. We compared importance ratings among all design features with Friedman chi-square tests (χ²) and paired comparisons within domains with Wilcoxon signed rank tests (W). We compared ratings between younger and older participants and between participants grouped by experience with technology and smoking frequency with Mann-Whitney U tests (U). We adjusted P values for post hoc comparisons using a Bonferroni correction. We computed statistics in R (version 2.15.1) [55].

Of the 119 respondents in total, 25 screened ineligible, one refused, and 38 could not be recontacted for eligibility screening. In all, 56 respondents were eligible and scheduled to attend a focus group stratified by age (<40 vs ≥40 years old); 16 failed to attend, leaving a final sample of 40 participants (Table 1). The younger focus groups included a total of 22 participants (P1-P22) and the older focus groups included a total of 18 participants (P23-P40). Participants were primarily white (63%, 25/40), non-Hispanic/Latino (90%, 36/40), and had less than a college degree (65%, 26/40). On average, participants smoked 12 cigarettes per day. Compared with younger participants, a higher proportion of older participants were heavy smokers. More than half of participants reported being “very experienced” with computers or better (58%, 23/40) and approximately half had health apps on their mobile phones (48%, 19/40). Examples of apps that emerged during group discussion focused largely on diet and fitness (eg, MyFitnessPal). Only five participants (13%) had downloaded a quit-smoking app before. Of the remaining 35 participants, the majority (83%, 29/35) reported they would consider downloading one.

Blue-sky brainstorming led to a range of design features smokers envisioned for future mHealth quit-smoking tools. Most participants expressed a strong overarching expectation for a highly individualized and responsive experience: “What if you had a profile where you could personalize it and make it your own?” (P1). Rather than follow a prescribed and impersonal quit process that several participants experienced with current apps, P12 wanted the ability to “...personalize your notification or whatever you want your phone to tell you. Write it yourself; set your own goals.” More specifically, the most universal design preferences that emerged across groups were personalized tracking with adaptive feedback and real-time peer support. Both of these themes reflect the critical importance participants placed on support to combat smoking cravings.

Participants’ preferred design options expanded on themes from their blue-sky ideation of mHealth quit-smoking tools. Their feedback on preferred options for messaging, social support, and gamification features provided deeper insight into their needs and priorities as target users.

Following focus group discussion that covered a wide range of design ideas and options, participants completed the exit survey to individually rate the importance of 21 specific design features (Table 2), each chosen to reflect best-practice treatment, ways to leverage mobile technology to support quitting smoking, or other important user preferences across domains of cost and reputation, privacy and security, content and user experience, and communication [27].

Participants rated some design features as significantly more important than others (χ²=189, P<.001). The highest ratings (n=40) were for tools that matched content to ones needs and interests, helped with withdrawal and side effects, and tracked progress. The lowest rated feature for all was posting to social media. This pattern was consistent across age groups, with the exception of older participants placing the greatest importance on privacy.

For readers interested in a granular examination of how perspectives on the 21 surveyed design features differed, we include a full list of paired comparisons within domains by each age group in the Multimedia Appendix 1. Briefly, importance ratings differed significantly within each domain (cost and reputation: χ²₃=21.2, P<.001; privacy and security: χ²₂=11.6, P=.003; content and user experience: χ²₈=96.1, P<.001; communication: χ²₄=46.8, P<.001). Paired comparisons suggested participants prioritized cost, maintaining privacy, and the ability to track, obtain personalized content, and support for nicotine withdrawal over other design features in mHealth tools. Participants also prioritized communicating about smoking outside of social media, perhaps with individuals whom are less familiar. For example, ratings were significantly higher for communicating with stop-smoking experts than family and friends (W=38, P=.03). Further, several participants shared the sentiment expressed by P32: “It’s a lot easier to be honest with somebody that you don’t personally know because like my doctor I’ve had her for 15 years and I love her, yet there’s stuff I don’t want her to know. But I want to be able to be honest at the same time.”

Compared with younger participants, older participants rated the following features as significantly more important: a tool that “stores information on your phone” (U=116, P=.02), “sends supportive or motivational messages” (U=113, P=.02), and “lets you communicate with stop-smoking experts about your progress” (U=122, P=.03). Compared with participants who had some or intermediate experience with technology (n=17), those who were very experienced or expert (n=23) rated “helps you track your progress” significantly more important (U=113, P=.01).

Compared with heavier smokers (n=23), lighter smokers (n=17) rated the importance of most features higher overall. Lighter smokers rated the following features significantly higher than heavier smokers: a tool that “matches content to your personal needs and interests” (U=250, P=.03) and “changes content as your needs and interests change” (U=271, P=.03). However, both of these features were among the features heavier smokers rated highest. There were no significant differences in importance of features rated by participants who were actively trying or thinking of quitting in the next 30 days (n=19) and participants thinking about quitting in 6 months (n=19).

Finally, after synthesizing the feedback from both the qualitative focus group feedback and qualitative survey results, several key design considerations emerged and are presented in Table 3.

Using a mixed-methods approach, we explored smokers’ ideal design features for mHealth cessation tools and assessed the relative importance of specific design considerations and treatment components that are particularly relevant to this design space. Our qualitative findings illuminated a number of creative design ideas, but most importantly, highlighted smokers’ desire for a highly personalized and adaptive experience, the ability to connect with peers for support, and the use of active and passive communication channels. Qualitative insights also highlighted preference for pull messages (ie, delivered upon request) over push messages (ie, sent automatically), interaction with other smokers through closed social networks, and targeted incentives based on user group. Quantitative results reinforce considerations of importance, including personalized content, the ability to track progress, support for nicotine withdrawal, connecting with peers and quit-smoking experts, and keeping information private. However, some preferences vary by age, experience with technology, or smoking frequency. For example, older smokers were more sensitized to issues of privacy and younger smokers were more interested in gaming as a means of smoking distraction. Smokers who had more experience with technology placed more importance on tracking features than those with less experience. Compared with heavy smokers, lighter smokers placed more importance on personalized content that matches and dynamically adapts to one’s changing needs and interest.

Although incorporating smokers’ preferences will not guarantee mHealth tools will be effective, balancing user preferences with best-practice treatment considerations could enhance program adoption and improve treatment outcomes. Thus, these findings offer guidance to addiction treatment experts and developers working in this design space.

The results of this study have implications for future research and development. For example, smokers wanted a highly personalized experience that included tracking (of smoking and smoking locations) and adaptively tailored content. In the future, tracking tools could couple self-monitoring [29] with passive sensing via wearable devices [56,57] and machine learning to predict smoking triggers and proactively intervene with real-time tailored recommendations [58]. This intelligent tracking should coincide with smokers’ cravings and deliver support when and where it is needed [59]. Notably, this type of passive push intervention is in contrast to smokers’ stated preference for active pull messaging to control timing on demand [28,31]. Remaining unanswered empirical questions include which type of intervention smokers would actually prefer based on real-world experience and how effective each strategy is relative to the other. These are important issues for future research.

Another design consideration worthy of further exploration is the use of gaming features in mHealth cessation programs. Relative to other design features, few studies to date have employed gamification strategies to help people quit smoking [24,38,52], but one could envision these strategies being used to incentivize participation or reaching milestones (eg, 48 hours smoke-free) using earned badges or offering game play to distract smokers during cravings to smoke. Among our participants, older smokers were more responsive to traditional rewards and incentives [29,60,61], whereas younger smokers were more open to gaming features. The use of gaming features have also been recommended by design experts [52,62], but further research is needed to inform whether these features can truly enhance treatment participation or outcome and for whom.

Social support from peers was also highly valued by our participants, as was support from stop-smoking experts. Both of these are recognized as important best-practice treatment components [7], so understanding how to most effectively integrate this support in mHealth interventions is another important consideration for future research. Our findings suggest that smokers want support from peers “on demand” and most preferred interactive modes of support over prerecorded peer testimonials, such as automated video messaging found ineffective for cessation in prior work [37]. Participants’ lack of interest in posting to social media echoes patients’ preferences for sharing personal health information on closed online networks that are condition-specific rather than general-purpose social media they regularly use [63]. However, with experience, smokers might find value posting in quit-smoking communities [64-66], particularly if combined with interventions tailored to one’s readiness to quit [67]. Open research questions remain about protecting privacy [64,66,68] while facilitating exchanges among social ties through which smoking cessation information is known to spread [39-42].

Our results further suggest communication with treatment experts may be more highly valued than sharing one’s progress with friends, family, or even a personal doctor. Congruent with prior research [29], these findings may reflect a preference for the anonymity of weak social ties (ie, peer smokers, treatment experts) over strong social ties. If so, this preference for greater anonymity would support the notion that smokers are “ambivalent socializers” who are simultaneously keen yet reluctant to engage with others about smoking via social media [68]. Yet this preference could create a treatment challenge. Support provided by a trusted clinician through counseling is an important element of best-practice cognitive behavioral-based nicotine dependence treatment [7], in part because it holds smokers accountable. Interacting with others anonymously could reduce the risk of peer pressure or embarrassment due to failures to quit [64], but could also reduce the effectiveness of the intervention due to lack of accountability. This raises yet another important area for future research to determine the effectiveness of anonymous peer support and expert advice provided in the context of mHealth cessation interventions.

Strengths of this study include a diverse demographic group and our mixed-method research design. The use of blue-sky thinking to illuminate smokers’ “perfect world” design priorities is a unique strength of this paper. To our knowledge, this approach has not been used previously in this context. It revealed many creative design ideas, some of which were further confirmed as important in participants’ quantitative survey ratings. Finally, our comparison of younger versus older smoker preferences highlighted important distinctions in these groups and is another strength of the study design.

The study also has some limitations. First, the findings may not generalize to all smokers due to the small sample size. Participants varied in age, education, experience with technology, and smoking frequency, yet their limited ethnic and racial diversity may have not captured opinions across known differences in smoking patterns [69]. It is also possible that the views represented do not generalize to smokers who do not own mobile devices, although the opinions of that group are less relevant in this context.

Finally, our results only reflect the perceived preferences of smokers’ for mHealth cessation tools. Future research is needed to determine if these stated preferences hold true in real-world practice. Such studies could lay the foundation for future work to assess the relative value and effectiveness of mHealth tools compared with other quit-smoking modalities (eg, peer support groups, professional behavioral support, pharmaceuticals, and nicotine replacement).

We assessed smokers’ design preferences for future mHealth smoking cessation tools and identified several features, which should be prioritized when designing future mHealth cessation tools. These include making programs that are highly personalized, adaptive, interactive, and can facilitate communication with peers and experts. Prioritizing these features in future mHealth interventions could improve their acceptability to smokers and program engagement, and improve treatment effectiveness as a result. However, the integration of popular mHealth features alone may not ensure a program’s use or effectiveness [32]. Future programs must also be grounded in relevant behavioral theory [59] and evidence-based treatment recommendations [13-15]. Aligning design preferences with evidence-based solutions is particularly important for addictive behaviors such as smoking [70] in which design priorities of treatment experts and smokers can differ [27].

Future mHealth smoking cessation tools will also require empirical validation. We highlight several important research questions worthy of investigation. It is our hope that this work will inform and inspire collaboration between addiction treatment experts, designers, and developers to create the next generation of mHealth smoking cessation tools.