Rationale, design, and methods for the development of a youth adapted Brief Tobacco Intervention plus automated text messaging for high school students

Abstract

Background:

Tobacco use is increasing among youth in the U.S. The current study seeks to develop and pilot test a universal group-based Youth Brief Tobacco Intervention (Y-BTI) plus mobile phone automated text messaging (ATM) for 9^th grade students to both prevent initiation among non-users and promote cessation among current users.

Methods:

This study will use a sequential, multi-method research design beginning with mixed methods formative work with 9^th grade students to adapt the existing young adult Brief Tobacco Intervention (BTI) for youth and develop automated text messages. The formative work with students will identify salient themes and strategies and optimal delivery schedules for the ATM for the interventions. The second phase of the study evaluates the Y-BTI and ATM through a pilot cluster randomized controlled trial that compares four treatment combinations: (1) Y-BTI + ATM, (2) Y-BTI alone, (3) ATM alone, or (4) standard of care. The Y-BTI is a single session, group-based intervention delivered in schools. The ATM intervention will provide comparable content to the Y-BTI but be delivered via text messaging 3–5 times per week for four weeks.

Conclusion:

In order to curb the rise of tobacco use among youth, interventions that are easily implemented and disseminated need to be developed. We aim to build upon previous research by showing that a universal group-based Y-BTI and ATM are effective in reducing tobacco use among a 9^th graders by preventing initiation among tobacco naïve youth and promoting cessation among current users.

Introduction

In 2020, 23.6% of high school students and 6.7% of middle school students reported past 30-day use of a tobacco product in the United States (U.S.).¹ Further, 22.5% of 10^th and 12^th graders reported vaping or e-cigarette use in the past month, leading the U.S. Surgeon General to declare youth vaping a public health epidemic.² While e-cigarettes are the most commonly reported tobacco product, used by 5.3 million youth, other tobacco products are frequently used including cigars (5.3%), cigarettes (4.3%), smokeless tobacco (3.5%), hookah (2.6%), and pipe tobacco (<1.0%).³ Roughly 90% of daily smokers started before the age of 18, and 2,000 youth smoke a cigarette for the first time each day in the US.^4,5 Further, e-cigarette use is on the rise among youth, and is linked to cigarette initiation among tobacco naïve youth.^6–8

Cessation at a younger age is associated with better health outcomes and lower mortality.⁹ However, youth who use multiple products report higher nicotine dependence and greater difficulty quitting.⁴ Further, while many young people who initiate tobacco use want to quit within a short time of commencing,^10,11 evidence suggests that nicotine addiction occurs very quickly in youth and young adults, making the odds of a successful quit attempt low.¹² Given the obligatory nature of school attendance, and the long history of students using tobacco in schools,¹³ schools have been the primary context for delivering health interventions targeting adolescents.^14,15 As such, there is substantial empirical evidence demonstrating long-term positive effects of school-based prevention programs in reducing adolescent tobacco use,^16,17 and a number of effective prevention programs have been made available for wide-scale dissemination. Unfortunately, the evidence for the long-term effectiveness of tobacco cessation programs for youth is much more limited.^18,19 In a recent Cochrane review of 41 clinical trials evaluating smoking cessation programs for youth, there was promising support for group counseling (9 studies, risk ratio [RR] 1.35, 95% confidence interval 1.03 to 1.77) to increase youth cessation, however automated text messaging was not effective (3 studies, RR = 1.18, 95% confidence interval 0.90 to 1.56).¹⁸ The authors concluded that they could not identify a smoking cessation program for youth that was superior to quitting unaided.¹⁸ Additionally, the majority of evidence-based tobacco control programs consist of 8–12 sessions,²⁰ competing for the limited time teachers have to meet educational requirements.²¹ Given the advent of stealth vaping, or vaping that can be used in schools discretely without drawing attention from teachers and administrators, it is critical that effective school-based tobacco cessation programs are developed, addressing both the wide range of currently available tobacco products and contemporary patterns of use, with implementation and sustainability in mind.

The Current Study

The current study seeks to fill this gap by adapting a universal group-based Youth Brief Tobacco Intervention (Y-BTI) and developing a mobile phone automated text messaging (ATM) for 9^th grade students enrolled in high schools in south and southwest Virginia. The goal of the study is to pilot test these interventions to determine their efficacy in both preventing non-users from initiating tobacco use as well as promoting cessation among current users. The Y-BTI is based on the Brief Tobacco Intervention (BTI) that has demonstrated efficacy in helping 18–20-year-olds reduce their tobacco use in a randomized clinical trial with young adults undergoing technical training in the military.²² The BTI is a single session, 45-minute group-based intervention, that uses principles of motivational interviewing to increase tobacco cessation among current users and prevent tobacco initiation among tobacco naïve individuals.²² Open-ended questions, reflections, and decisional balance are used to promote a desire to remain tobacco free.²³ Additional details regarding the formative assessment process of the BTI can be found in Little et al.²⁴

There have only been three randomized clinical trials evaluating the effects of ATM to increase tobacco cessation among youth. While the risk ratio across these cessation studies demonstrated that these interventions were not effective in promoting tobacco abstinence on their own (three studies, RR = 1.18, 95% confidence interval = 0.90 to 1.56),¹⁸ ATM interventions are effective in promoting tobacco cessation with adults,²⁵ suggesting that the right combination of messages could prevent and reduce tobacco use among youth. Additionally, we could not find any studies testing the use of ATM to prevent tobacco initiation among youth.

However, there are several advantages to using ATM, including the ubiquity of cellphone ownership (88% of 13- to 18-year-olds own a cellphone),²⁶ particularly among adolescents, and their comfort interacting with a cellphone via text messaging. Given that all content is set and planned in advance, confidentiality is maintained while ensuring that all messages are received by participants. Lastly, if effective, it is easy to scale since individual counselors are not needed to implement the intervention, and the intervention does not impede on instructional time. Therefore, the ATM affords a cost-effective and scalable alternative or complement to the group-based intervention.

In the current paper, we describe the design of an ongoing trial to adapt the previously validated BTI for youth and develop an ATM to prevent and reduce tobacco use among 9^th grade youth. Given the universal nature of the proposed study, 9^th grade was chosen as many youths initiate in middle school, but the majority of initiation occurs during high school. In 2019, 33.5% of 9^th graders in Southwestern Virginia reported currently smoking cigarettes or electronic cigarettes.²⁷ Therefore, 9^th grade is a logical time to prevent initiation among non-users, while also promoting cessation among those who have already started using tobacco in this region.

Materials and Methods

Study Design

This study will use a sequential, multi-method research design including (1) mixed-methods formative work with students, (2) a pilot of the intervention content with two classes of students per school, and (3) a cluster randomized controlled trial of the Y-BTI and ATM interventions. The mixed-methods formative work includes a quantitative survey assessment as well as a qualitative sticky note activity with whole group discussions with students to identify salient intervention themes and strategies for targeting the Y-BTI intervention and developing the text messages. The pilot test will evaluate the content and themes present in the interventions to ensure that they are credible and believable by students. The second phase of the study evaluates the Y-BTI and ATM through a pilot cluster randomized controlled trial that compares four treatment combinations: (1) Y-BTI + ATM, (2) Y-BTI alone, (3) ATM alone, or (4) standard of care. The primary goal of the randomized controlled trial is to obtain effect size estimates for each treatment intervention to accurately power a larger clinical trial to evaluate the long-term effectiveness of the interventions. The Y-BTI + ATM will consist of an in-person group-based brief tobacco intervention and three - five automated text messages per week for four weeks. In the standard of care group, students will only receive the tobacco prevention or cessation activities, if any, offered directly by their school. The primary outcome in Phase 2 will be past 30-day abstinence²⁸ from all tobacco products 28 weeks after enrollment. Participating schools will come from three school districts located in South- and South-Central regions of Virginia, where there are high rates of tobacco use among youth.²⁷ The school districts were chosen based on existing relationships in the regions with the University of Virginia Cancer Center’s Office of Community Outreach and Engagement. This study was approved by the Social Behavioral Sciences Institutional Review Board at the University of Virginia.

Phase 1: Mixed-Methods Formative Work.

Participants and Procedure.

We will conduct a quantitative survey assessment as well as a qualitative sticky note activity with whole group discussions with approximately eight 9^th grade health classes (n=216) enrolled at participating schools in Virginia. No personal identifying information will be obtained from study participants, and thus an opt-out or passive consent process will be used.

Two weeks prior to the formative assessments, staff will visit each of the pre-selected 9th grade classes and describe the study protocol, answer questions, and discuss the opt-out form. Students will be instructed to discuss the study with their parents or legal guardians and return the form to their teacher if their parents or guardians do not want them to participate.

Two weeks later, staff will return to the same pre-selected 9th grade classes to conduct a quantitative survey assessment and the sticky note activity with group discussions. Students whose parents opted them out will be sent to another classroom. Students who participate in the activity will receive a $5 gift card for participating following the activity.

Mixed-Methods Formative Work Methodology.

At the start of the session discussion, students will be asked to complete an anonymous paper-and-pencil survey^24,29 to collect data from all participants, including participants who may not feel comfortable expressing their opinions out loud during the group discussions. The quantitative assessment will measure 1) demographics (e.g., sex, race, ethnicity), 2) past use of tobacco, 3) intentions to use tobacco, 4) willingness to use tobacco, 5) peer use, 6) perceived harm of tobacco products, 7) perceived addictiveness of tobacco products, 8) text message evaluations (e.g., how believable, credible, interesting), and 9) opinions about acceptability and timing of different text message delivery schedules. For a detailed description of the measures see the Study Measures section.

After completing the quantitative survey, students will be provided with a pack of sticky notes and asked six open-ended questions. Question topics will include: a) facilitators and deterrents of tobacco product use among youth, b) reactions to pro- and anti-tobacco messages, c) perceptions of tobacco addiction and harm, and d) ideas for framing and introducing the interventions to make messages more relevant to students. After each question, students will be asked to write out their responses and place their sticky notes on a poster sized paper that will be turned around so it is not facing students to offer great anonymity as students place their sticky notes on the board. After students have answered the questions, the facilitator will lead the class in a group discussion, summarizing the responses on the sticky notes and facilitating additional discussion through the use of open-ended questions. The whole activity takes about 30 minutes. Following the discussion, posters will be photographed for documentation. Study staff will also take notes during discussions to note major themes. Prior to launching any of the in-school activities, staff will undergo training on the study protocol, data collection procedures and notetaking to ensure that uniform procedures were followed.

Mixed-Methods Formative Work Data Analysis.

Data from the survey assessment will be described using proportions, means with standard deviations, and medians with interquartile ranges as appropriate. Differences will be calculated and compared using chi-square tests for categorical variables, and non-parametric tests for more than two samples. Significance will be set at p< 0.05.

Photographed responses from the sticky note discussions and notes from the qualitative discussions will be transcribed in Nvivo software using a thematic analysis approach,^31–33 allowing us to discover emerging perceptions and meaningful categories and to iteratively generate themes during the coding process. Two independent researchers will generate initial codes and meet to discuss any discrepancies. A third reviewer will review all initial codes by the two independent researchers, resolve discrepancies that could not be reached between the two researchers, finalize codes, and develop a codebook with definitions. Through this process we will explore common and contradictory themes across groups. Based on the findings from the sticky note activity and discussions, we will select the messages and delivery schedules for the ATM. Descriptive statistics will be calculated from the anonymous survey; survey responses, the sticky note activity, and qualitative discussion data will be triangulated to strengthen findings.³⁴

Phase 1: Pilot Test.

Participants and Procedures.

Prior to the randomized controlled trial, we will pilot test the adapted Y-BTI with approximately eight classes of students (n= 216) enrolled in 9th grade health in each of our participating high schools. Identical to the procedures for the sticky note activity and discussions, no personal identifying information will be obtained from study participants, and an opt-out or passive consent process will be used. On the date of the pilot test, the students whose parents opted them out will be sent to another classroom for the duration of the class period. Students will receive a $5 gift card for participating.

Finalizing and Piloting the Interventions.

Students will complete the anonymous paper-and-pencil survey used in the sticky note activity and discussions, which will be updated to include the proposed text messages in the ATM intervention to determine student perceptions related to believability, credibility, and interest. Next, staff will deliver the Y-BTI, followed by a class discussion on the acceptability and perceived effectiveness of the intervention (likelihood that the interventions messages will lead them to reduce or avoid tobacco).³⁵

Phase 2: Randomized Controlled Trial.

Study Design.

This study is conceptualized as a two-factor group randomized controlled trial with four treatment combinations: 1) Y-BTI + ATM, 2) Y-BTI alone, 3) ATM alone, or 4) standard of care. The Y-BTI treatment condition will include the classroom-based 45-minute in-person intervention. The ATM intervention will last four weeks. The combination treatment will receive both the Y-BTI and the ATM intervention. The control condition will only complete assessments. Due to the potential for cross contamination between students, randomization will occur at the school level and intraclass correlation coefficient (ICC) will be used to measure similarity between participants within a school. We will randomize eight high schools to one of each condition.

The independent variable will be treatment assignment and the main outcome variable will be past 30-day abstinence²⁸ at the 28-week follow-up. Secondary outcomes are seven-day point prevalence abstinence from all tobacco products at the 28-week follow-up. Figure 1 shows the four unique combinations of factors in the proposed study. Process measures (e.g., fidelity of intervention, dosage of text messages) will be obtained throughout the trial.

Figure 1.

Design of the Randomized Clinical Trial.

Note. Students receiving the ATM intervention will sign up the same day as the Y-BTI. Y-BTI = brief tobacco intervention, ATM = automatic text messaging, FU = follow-up survey, Wk = week).

Theoretical Framework

Our theoretical framework draws from the Theory of Planned Behavior,³⁶ social diffusion model of campaign influences,^37,38and elements from Behavioral Economic Theory.³⁹ We propose that the combination of the group-based intervention plus ATM will impact tobacco use directly by changing attitudes, subjective norms, and perceived behavioral control among people who receive the messages. The interventions may also have an indirect influence by promoting interpersonal conversations between students. Conversations about the messages between students within the same social networks may generate more in-depth thinking about the intervention messages, leading to changes in attitudes, social norms, and perceived behavioral control. As a result, these perceptions may change behavioral intentions, which may lead to changes in tobacco use behavior. The Theory of Planned Behavior suggests that interventions focus on shaping one’s attitudes, subjective norms, or perceptions of behavioral control in order to produce a desired behavior change.⁴⁰ Guided by the Theory of Planned Behavior, the interventions will have 3 targets: (1) fostering negative attitudes towards tobacco; (2) changing subjective norms (denormalizing tobacco use); and (3) enhancing perceived behavioral control (self-efficacy to remain tobacco-free).

Randomized Interventions

Youth-Brief Tobacco Intervention (Y-BTI).

The universal prevention program was designed as a group intervention to include components of effective tobacco control programs for youth and young adults.⁴¹ The intervention lasts approximately 45 minutes and is delivered in a classroom setting, utilizing a Socratic teaching style and evoking participation using the principles of motivational interviewing.²³ The Y-BTI targets all tobacco products with the goals to enhance motivation for youth to quit tobacco or remain tobacco-free, reduce intentions to use tobacco, promote peer discussions around the impact of using tobacco, and correct cognitive misperceptions around tobacco use (see Table 1). A detailed description of the BTI can be found in the work of Little et al.^22,24,42 Based on the information gained regarding prevalence, attitudes, and beliefs about the various forms of tobacco from Phase 1 of the project, we will be able to tailor the focus of the interventions accordingly, spending more time on the tobacco products that are used more frequently by students.

Table 1.

Examples of Active Components of the Interventions

Intervention Target	Strategy	Sample Active Component
		Y-BTI	ATM
Perceived Behavioral Control	Self-efficacy to remain tobacco-free	-Enhance refusal self-efficacy skills by role playing situations where students might feel pressured to use tobacco and act out how students could refuse	“Think about what you might say if someone offered you their JUUL.” This message would be followed by, “Did you think about what you could say to someone who offered you their JUUL? Text YES or NO.” If YES, “Great! Write that idea down somewhere so you can look back at it if needed”. If NO, “That’s okay. Here is one idea you could use: Thanks, but I’m not interested in trying that today”.
Subjective Norms	Correct cognitive misperceptions	-Videos depicting peer groups expressing anti-tobacco sentiment	“How many kids in your class do you think dip or chew?” This message would be followed by, “It’s actually X%. Was your guess higher or lower than the actual number?” Similar messages would be sent to denormalize other tobacco product use.
Attitudes towards tobacco use	Decisional Balance	-Student led discussion of the positive and negative aspects of using tobacco	“Some kids think vaping is harmless. Check out this fact sheet and decide for yourself: https://bit.ly/2Y7bJkL”
Behavioral Economics Approach	Focus on tobacco-free reinforcing activities	Students identify enjoyable and goal-consistent tobacco-free activities they could engage in during their free time (e.g., sports/exercise, online classes, reading, movies)	“Identify a long-term goal and one step you can do this week to move closer to that goal.” This message would be followed by, “Did you think about what you might do this week to get closer to reaching your goal? How might using tobacco this week get in the way?”

Note. Y-BTI = Brief Tobacco Intervention; ATM = Automated Text Messaging

ATM Intervention.

The ATM intervention is automated because all messages are pre-written, and their timing is pre-planned. Therefore, all students will receive the same content at the same time.

The ATM intervention will last four weeks, with three - five messages sent per week. The automated program will include a mix of static and responsive messages. Specifically, some messages will be designed to maintain or enhance knowledge about tobacco use by providing facts to the student, for example, listing the negative health effects of tobacco use. Other messages will promote engagement and reflection from the student. To promote refection, some messages will request a response, but in a way that maintains confidentiality. The text messages will extend the constructs presented in the Y-BTI. Table 1 presents sample messages by intervention target showing overlap between the two interventions. To ensure confidential information is not sent to the text messaging program, messages that need a response will include the appropriate replies, while reflection questions will be left open-ended. Additionally, we have secured a Certificate of Confidentiality for the current study from the National Institutes of Health. Students enrolled in the ATM condition will have the option of opting out of receiving the messages by texting ‘STOP’. We will assess the frequency and timing of opt-out from the ATM intervention (e.g., texting “STOP”).

Data Collection and Assessments

Participants.

Within each participating high school, two health classes will be randomly selected to participate. We estimate that a total of 432 students will be enrolled in the classes selected for participation.

Procedure.

Trained study staff will visit each class and describe the study protocol, answer questions, and read through the consent form. Students will be instructed to discuss the study with their parents or legal guardians and return the signed consent form if they are interested in participating. Students who receive positive parental consent will be asked to provide assent prior to taking the baseline survey. Students will receive a $5 gift card for bringing back a signed consent form.

Baseline and follow-up measures will be collected from students using a self-report, closed-ended response pen and paper survey. At each wave of data collection, surveys will be administered over one class period by the research project staff to students with positive parental consent and student assent. Those who consented to participate but are absent on assessment days will be left absentee packets, and school staff will be asked to distribute the surveys to the students and return the completed surveys in a pre-stamped and addressed envelope.

Following completion of the baseline assessment, students within schools randomized to the Y-BTI alone or Y-BTI + ATM will receive the Y-BTI. It is important to point out that the school districts in the region strongly support tobacco control education. Therefore, the Y-BTI component of the intervention will be a required intervention for any students at the schools randomized to receive the Y-BTI, while participation in this study would be completely voluntary.

Students from schools randomized to receive the ATM alone or Y-BTI + ATM will begin receiving the text messages within a few days of completing the baseline survey. We anicipate running the ATM program for four weeks, with on average three - five messages sent per week. However, these details might change based on the formative work completed in Phase 1 of the study. For schools randomized to receive the ATM, consented students will be directed to enroll into the supplemental program by texting an entry code to a dedicated study short-code (five - six numbers that can be used to send and receive text messages). The study staff will be available to assist students in registration to the ATM intervention. If a student does not have a cell phone, they will not be enrolled in the ATM intervention, and it will be noted in their study record.

To assess treatment fidelity we will utilize a checklist and assess intervention delivery of all Y-BTI sessions. Treatment fidelity will be entered in the analytic model as a potential covariate in secondary data analysis. In addition, prior to implmentation of the Y-BTI we will have an all-day training on delivery of the interventions, background information and the opportunity to practice delivery prior to launch. Interventionists will also participate in weekly study meetings to review overall adherence to structured protocols and prevent any drift between the two interventionists. These study meetings (attended by all investigators and the interventionists) will also include problem solving related to intervention delivery.

For all study participants, 28 weeks after the baseline survey is completed, trained study staff will return to each class and complete the follow-up assessment with students. A 28-week follow-up was chosen since it is 6 months after the end of the ATM intervention for those who receive that condition.

Study Measures.

Participants will complete a questionnaire at baseline and then again at a 28-week follow-up. Baseline and follow-up questionnaires will contain the same measures, with the exception that demographics will only be assessed at baseline, while point prevalence abstinence, continuous abstinence and program satisfaction will only be assessed at follow-up.

Demographic Measures.

Demographic items include sex, age, ethnicity, race, zip code, socio-economic status, cellphone ownership, and internet access and use. Cellphone ownership will differentiate between owning a smartphone, basic cellphone and not having any cellphone at all. The participants will be asked to report how frequently they access the internet at home, school, in public, or on mobile devices on a scale ranging from never at this location to daily at this location.

Tobacco Use History, Point Prevalence and Continuous Abstinence.

Lifetime tobacco use (ever tried the following tobacco products, even once or twice; response options of yes, no, or don’t know/can’t remember), past 30-day use (response options of every day, some days, and not at all), point prevalence abstinence, continuous abstinence (from baseline to the 28-week follow-up) and quit attempts in the past six months will be measured for each product (cigarettes/roll your own, dip/smokeless tobacco/snus, e-cigarettes/vape, pipe filled with tobacco, cigarillos/little cigars, cigars, and hookah).^43,44

Attitudes and Perceptions.

Perceptions of harm and addictiveness of different tobacco products will be assessed on a three-point scale (response options of not at all harmful/addictive, a little harmful/addictive, very harmful/addictive, and I don’t know). Perceptions of the dangers, addictiveness, and potential harms of each product with answers on a five-point Likert scale from never to always.⁴⁵

Participants will rate their perceptions of each tobacco product on three different five-point scales: bad to good, harmful to beneficial, and foolish to wise. Students will also be asked “How much do the following characteristics describe a typical tobacco user,” and will be presented with a list of positive (clean, cool, sexy, smart, and healthy) and negative (trashy, disgusting, immature, unattractive and inconsiderate) traits to describe cigarette and e-cigarette users. Answer options for these characteristics included not at all, a little bit, somewhat, quite a bit, and very much.⁴⁶

Lastly, participants will be asked to rate the acceptability of certain behaviors (i.e., taking one puff, using for a while, or saving for later) if offered different tobacco products. Answer options will be on a five-point Likert scale from unacceptable to acceptable.⁴⁷

Subjective Norms.

Descriptive norms will be assessed by asking participants how many students in their school they think use different tobacco products. Injunctive norms will measure the acceptability of their use of tobacco products by their closest friends and family members.⁴⁸ Peer use of tobacco products will be assessed by asking the participant how many of their five closest friends use any tobacco products. Participants will report how often anyone smokes, vapes, or uses dip or chew tobacco inside of their home at baseline, with answer options of daily, weekly, monthly, less than monthly, and never.

Perceived Behavioral Control, Susceptibility, and Intentions.

Participants will rate how confident they believe they can remain tobacco-free for the next six months on a five-point Likert scale from not at all confident to completely confident. They will also be asked how often they think important people in their life (e.g., friends and family) are confident in the student’s ability to remain tobacco free on a five-point Likert scale ranging from never to always. Participants will also be asked about their willingness to use tobacco products if offered: 1) take one puff, 2) use one for a while, 3) take on to try later. Items are on a five-point Likert scale ranging from not at all to extremely willing.⁴⁹ Finally, tobacco susceptibility will be asked using three items,⁵⁰ such as “Do you think that in the future you might experiment with any of the following tobacco products?” Responses will be on a four-point scale from definitely not to definitely yes.

To assess for behavioral intentions, participants will also be asked about their intentions to use any form of tobacco in the next 12 months using a 4-point Likert scale from definitely not to definitely yes.⁴⁸

Program Satisfaction.

Students will be asked about the acceptability of the intervention,⁵¹ and students who received text messages will be asked who they shared the messages with to determine if the messages spurred interpersonal conversations about tobacco use.^51,52 Students will also be asked if they discussed the content of each of the interventions with anyone from their school or another school.

Statistical Analyses and Power Calculations

Power Calculations.

The primary goal of the randomized controlled trial is to obtain effect size estimates for each treatment intervention to accurately power a larger clinical trial to evaluate the long-term effectiveness of the interventions. Despite that goal, we briefly discuss the minimum detectable overall abstinence rate based on the proposed sample.

All power and sample size calculations were completed with PASS12⁵³, and corroborated by simulation using data analysis models proposed in this study. The interventions are delivered to two classes of approximately 27 students each from the 8 participating high schools. Therefore, we must account for the clustering effect of students nested within classes, nested within schools. We will randomize eight schools (approximately N=432 9^th graders) to each study arm (N_{Y-BTI + ATM} = 108 9^th graders, N_Y-BTI= 108 9^th graders, N_ATM= 108 9^th graders, N_Control = 108 9^th graders). Assuming a conservative ICC of .01 among schools (based on previous clinical trial research⁵⁴) and control abstinence rate of 7% at 28 weeks, we estimated a minimally detectable difference of 18% (7% to 25%) between treatment groups with 80% power and alpha = .05, with 7% to 21% detectable as ICC approaches 0. After rounding, results are the same after admitting a 4% attrition rate (N=26 students per class).^18,55 When assuming conservative group abstinence rates, the simulation models run with 1,000 replicates suggested that the Y-BTI (assuming OR = 2.16) and the ATM+Y-BTI (assuming OR = 2.53) effect estimates will be greater than an OR = 1.50 (weak effect) for 78% and 89% of the time, respectively.

Statistical Analyses.

All analyses will be performed using SASv9.4. Examination of distributions may prompt transformations, where they are justifiable. We will perform analyses to determine whether differential attrition occurred by subjects’ baseline characteristics. Analysis will be intent-to-treat. All statistical tests will be two-sided.

We will describe baseline characteristics at the individual, class, and school level using proportions, means with standard deviations, and medians with interquartile ranges as appropriate across the four groups, and individual, class, and school level descriptive statistics related to baseline characteristics will be calculated and compared using chi-square tests for categorical variables, and non-parametric tests for more than two samples. If sample imbalance is either statistically detectable or minimally important, follow-up sensitivity analyses will be conducted by adjusting for imbalanced covariates in the generalized estimating equation regression.

To determine the efficacy of the interventions on tobacco abstinence, generalized estimating equation models for binary data⁵⁶ will be used to estimate 30 day tobacco abstinence at the 28-week follow-up, adjusting for indicators for each of the treatment factors and all two-way interactions between the treatment factors, while accounting for the structure of the data (i.e., students nested within classrooms within schools). Additionally, several demographic and smoking history variables will be considered to assess potential effect modification. We will also control for the potential of contamination across schools by including an item that asks students if they discussed an intervention that they were not randomly assigned to receive with their friends, and potential exposure to any school implemented tobacco prevention program.

Secondary analyses will include examining the effect of the interventions on e-cigarette use and continuous abstinence. We will apply the same model to test for continuous e-cigarette abstinence (coded yes/no) at follow-up. To determine whether the intervention had a prevention or cessation effect, or perhaps both, we will apply the same model to test the interaction between the treatment assignment and tobacco use history (tobacco use status at baseline coded yes/no).

Conclusions

The current study will use a sequential, multi-method research design to adapt an evidence-based tobacco prevention and control program for high school youth. We aim to build upon previous research showing that a universal group-based BTI is effective in reducing tobacco use among a large, racially diverse sample of young adults.⁴² Through this study, we will also develop an adjunct automated text messaging program that provides comparable content to the Y-BTI but delivered individually over a longer period (recall that the Y-BTI is a one session group-based intervention delivered in the classroom).

There is a dearth of tobacco prevention text-messaging interventions. Despite the fact that 91% of texts are read within the first 3 minutes of receipt (Gates et al., 2014)⁵⁷and the majority of high school students own cellphones,²⁶ most adolescent tobacco prevention programs are delivered via classroom group-based curriculum. The current study is novel in that it develops a youth prevention ATM intervention that can reach students without impeding on valuable instructional time. In the current study, we plan to send students randomized to the ATM condition 3–5 text messages per week since many of the students receiving the texts are not using tobacco or actively trying to quit. It will be critical that a balance in struck between reminding students of the intervention themes, while not overburdening them, leading to students opting out of the ATM component. These results will provide valuable insight into the feasibility and effectiveness of an ATM tobacco prevention program for youth.

However, the study is not without limitations. The current evaluation is a pilot study that is not powered to detect effects. If shown to be well-received and promising, then a large-scale evaluation of the study is warranted. The study is also limited to students from Virginia. As such, the effects of the intervention might not translate to youth in other regions of the country. Lastly, our outcome of interest is self-reported tobacco abstinence, and therefore may elicit social desirability bias. However, we do not expect that this bias would be differential by study arm.

Drawing from the Theory of Planned Behavior³⁶ and Behavioral Economic Theory,³⁹ we believe that the Y-BTI + plus ATM will effectively change youth attitudes and perceptions regarding tobacco while evoking conversations about tobacco among students, and ultimately reducing behavioral intentions to use tobacco and subsequent tobacco use behaviors. This is important given the changing tobacco landscape and that cessation at a younger age has been associated with better health outcomes and less mortality.⁹ Unfortunately, there is limited evidence for the long-term effectiveness of tobacco cessation programs for youth.^18,19 If the interventions in the current study are proven efficacious, they can easily be disseminated to other schools to continue reduce youth tobacco use.