Advertising Disclaimer »

Article

Effects of Physical Activity on Teen Smoking Cessation

Kimberly Horn, Geri Dino, Steven A. Branstetter, Jianjun Zhang, N Noerachmanto, Traci Jarrett and Melissa Taylor
Pediatrics October 2011, 128 (4) e801-e811; DOI: https://doi.org/10.1542/peds.2010-2599
Download PDF

Abstract

OBJECTIVE: To understand the influence of physical activity on teen smoking-cessation outcomes.

METHODS: Teens (N = 233; 14–19 years of age) from West Virginia high schools who smoked >1 cigarette in the previous 30 days were included. High schools with >300 students were selected randomly and assigned to brief intervention (BI), Not on Tobacco (N-O-T) (a proven teen cessation program), or N-O-T plus a physical activity module (N-O-T+FIT). Quit rates were determined 3 and 6 months after baseline by using self-classified and 7-day point prevalence quit rates, and carbon monoxide validation was obtained at the 3-month follow-up evaluation.

RESULTS: Trends for observed and imputed self-classified and 7-day point prevalence rates indicated that teens in the N-O-T+FIT group had significantly higher cessation rates compared with those in the N-O-T and BI groups. Effect sizes were large. Overall, girls quit more successfully with N-O-T compared with BI (relative risk [RR]: >∞) 3 months after baseline, and boys responded better to N-O-T+FIT than to BI (RR: 2–3) or to N-O-T (RR: 1–2). Youths in the N-O-T+FIT group, compared with those in the N-O-T group, had greater likelihood of cessation (RR: 1.48) at 6 months. The control group included an unusually large proportion of participants in the precontemplation stage at enrollment, but there were no significant differences in outcomes between BI and N-O-T (z = 0.94; P = .17) or N-O-T+FIT (z = 1.12; P = .13) participants in the precontemplation stage.

CONCLUSIONS: Adding physical activity to N-O-T may enhance cessation success, particularly among boys.

WHAT'S KNOWN ON THIS SUBJECT:

Studies of adult smokers have suggested that enhanced physical activity during attempts to quit may decrease withdrawal symptoms and cigarette cravings. Other studies have suggested that physical activity prescribed as part of adult smoking-cessation interventions may help sustain cessation over time.

WHAT THIS STUDY ADDS:

Adding physical activity to a youth smoking-cessation program may enhance cessation success, particularly among boys.

Although there is a growing body of literature on effective smoking-cessation interventions,1 there is limited knowledge regarding the factors that mediate cessation among teen smokers. Emerging literature suggests that physical activity, that is, bodily movements that enhance or maintain physical fitness and overall health, may mediate smoking cessation.2,–,4 Potential mediating mechanisms include reductions in weight gain, withdrawal symptoms, and cigarette cravings; notably, the latter 2 factors are known contributors to cessation resistance and smoking relapse.4,–,8 Studies on the relationship between physical activity and smoking among adults have reported both significant and nonsignificant findings. A systematic review of 13 randomized controlled trials by Ussher et al3 found mixed results for physical activity interventions as catalysts for smoking abstinence. Three studies showed significantly higher smoking abstinence rates among participants who received the physical activity intervention, compared with control subjects, at the end of the intervention period. At the 12-month observation, however, only 1 of the 13 trials showed significance for exercise as a predictor of smoking cessation. In contrast, an extended adult intervention study using pedometers and fitness counseling found that moderate/vigorous physical activity was positively associated with sustained smoking abstinence at 24 weeks.9 Interestingly, women had significantly higher quit rates in a cessation intervention, compared with programs with no fitness component.4 Ultimately, Ussher et al3 concluded that there is no harm in encouraging physical activity as a smoking-cessation aid, because there is no evidence that it interferes with cessation. Dual intervention efforts seem logical, because a majority of youth and adult smokers are physically inactive9 and a large body of evidence demonstrates the critical importance of physical activity for health throughout the life span.2

The research described above focused exclusively on adults, but there is evidence that physical activity, ranging from aerobic activities (eg, team sports) to less-vigorous activities (eg, walking), is protective against smoking initiation and acceleration among youths.10,–,12 Rodriguez et al13 posited that, when physical activity is prescribed for teens, it may be important to consider how youths interpret or perceive the experience. The authors found, for example, that the effects of interscholastic sports participation on adolescent smoking might be related to how competent teens feel after the activity. Those findings suggest that motivation and self-efficacy may be important considerations for implementation of physical activity adjuncts to smoking interventions. To our knowledge, no published research has examined the value of physical activity in conjunction with teen smoking-cessation efforts. Importantly, the teen years are a time when physical activity often decreases significantly; this may particularly true for teen smokers.14

The present study developed a physical activity component (FIT) as an adjunct for an evidence-based teen smoking intervention, Not on Tobacco (N-O-T).15 We used a randomized group design with 3 conditions, that is, brief intervention (BI) versus N-O-T versus N-O-T+FIT. We hypothesized that youths who received N-O-T+FIT would show significantly higher smoking-cessation rates than those who received N-O-T or BI. Consistent with previous N-O-T research,16 we also predicted that youths who received any exposure to N-O-T (whether as N-O-T or as N-O-T+FIT) would have higher quit rates than their counterparts who received BI. N-O-T was selected as the program of use because it has been demonstrated to be effective and cost-efficient.17,–,19 The program has been adopted by the American Lung Association and has received several federal designations.20 Enhancing the potential applicability of our findings, a study found N-O-T to be the most widely used teen smoking intervention in the nation.21

METHODS

Participants

Participants were enrolled from West Virginia public high schools between 2006 and 2009. The study included teens (14–19 years of age) who volunteered to participate. Our recruitment emphasized daily smokers but, given the variability in how teens perceive smoking status, we maintained a flexible inclusion criterion of ≥1 day of smoking in the previous 30 days, which is a well-established standard definition of a current smoker.22 Study procedures received West Virginia University institutional review board approval, with active parental consent and youth assent.

Intervention Conditions

Table 1 describes the study conditions for N-O-T+FIT, N-O-T, and BI. Extensive N-O-T program descriptions are available elsewhere.23,24,–,26 Although the core N-O-T curriculum includes content related to healthy lifestyles, such as generally increasing physical activity, improving nutrition, managing stress, and obtaining social support, the physical activity module (FIT) is more detailed, theoretically and empirically driven, gender-tailored, and linked specifically to N-O-T on a session-by-session basis.

View this table:
TABLE 1

Description of Study Conditions

Study Design

Of the 123 high schools in West Virginia, 99 public schools met the inclusion criterion of having ≥300 enrolled students. Randomly selected schools were assigned randomly to 1 of the 3 study conditions (BI, N-O-T, or NOT+FIT) by using an assigned code in a SPSS database (SPSS Inc, Chicago, IL) (Fig 1). Of 60 schools that were selected randomly in 3 waves, 40 agreed to participate; schools were then assigned randomly. After random assignment but before study onset, 21 schools dropped out, citing recruitment challenges and time constraints. Schools dropped out equally across conditions, leaving a total of 19 schools. Our final teen sample included 233 participants.

FIGURE 1
FIGURE 1

Random selection and assignment process. WV indicates West Virginia.

Facilitator Training

School staff members and/or principals helped to identify facilitators across conditions. The research team and American Lung Association West Virginia provided training to facilitators (N = 25) according to condition. According to the American Lung Association protocol, N-O-T facilitator training occurred over a 1-day (8-hour) period; an additional half-day (4 hours) focused on the research protocol.27 Training topics included teen smoking and nicotine dependence, research basics, and National Institutes of Health ethics certification. The BI training lasted <3 hours and focused primarily on the research protocol. Trained facilitators initiated recruitment in their respective schools and provided interested students with the institutional review board–approved consent/assent forms, to be signed by guardians and teens and returned before enrollment in any condition (Table 2).

View this table:
TABLE 2

Targeted Individual Enrollment

Data Sources

Researchers collected participant baseline data before the onset of intervention (±2 weeks). Follow-up evaluations occurred 3 and 6 months after the baseline assessment (±3 weeks), equivalent to 1 and 4 months after the quit date, respectively, for all conditions. In this article, we describe only measures relevant to the primary hypothesis for quit outcomes. Table 3 provides a summary of the 18 key variables used to assess the participants' baseline similarity, selected on the basis of their potential relationships to our primary outcomes.15,19,28,29

View this table:
TABLE 3

Key Variables for Baseline Comparisons

In addition to collecting data on daily cigarette use at the follow-up evaluations, we asked participants to classify themselves as “quit” or “not quit.” Smoking abstinence was verified with exhaled carbon monoxide (CO) readings, which are valid for a few hours. A digital CO monitor provides a reading of CO levels in the body. As in previous N-O-T studies and other teen smoking studies,30,31 our study applied a CO cutoff reading of <9 ppm to corroborate self-reports at the 3-month follow-up evaluation.

Analyses

The appropriate definition of cessation is subject to debate, especially in youth cessation studies.32,–,34 In alignment with emerging field standards,31 we examined 2 categories of quitting, that is, self-classified point prevalence (the proportion of teens who reported being a quitter at the assessment, regardless of the number of days since the last cigarette) and 7-day point prevalence (the proportion of teens who reported having no cigarette use in ≥7 days at the assessment). Both criteria used the CO-validated, intent-to-treat sample with both observed and imputed data (described below).

Little attention has been given to missing data in published teen smoking-cessation studies. Most studies report outcomes by using intent-to-treat or compliant (ie, completers) subsample analyses and view any missing data as “failure.” Research shows that missing values in cessation trials can lead to biased estimates of outcomes that do not represent accurately the potential enrolled population.16 Popular methods for dealing with missing data to produce complete data sets have serious drawbacks.35 List-wise deletion can introduce biased estimates and reduced power, and simple mean replacement leads to reduction of the SE estimates. Similar to procedures used by Joffe et al16 in a 2-armed randomized trial of N-O-T, the multiple-imputation procedure described by Rubin36 replaces missing values with a set of plausible values, eliminating uncertainty about which are the correct values to impute. The procedure introduces “error” while maintaining adequate power.

An important step in multiple imputation is determining the “missingness” of the data, particularly data used to calculate the key outcomes. Three months after baseline, we obtained data from 75% of our subjects (174 of 233 subjects). Rates of missing data for our 3 key outcome variables (quit status, CO reading, and days since the last cigarette) ranged between 25% and 39%. At 6 months after baseline, we obtained data from 63% of our subjects (146 subjects), with missing data rates of 37% to 74%. At 6 months after baseline, many youths had graduated and did not have the opportunity to complete the CO record in person (including the date of the last cigarette), which resulted in high rates of missing values. Therefore, we did not apply imputation, and we reported self-classified point prevalence only.

Consistent with past N-O-T research,15 reasons for missing data were not predictably related to smoking status (eg, transfer to a different school, need to be at work, illness, or graduation). Following the recommendations of Rubin,35 we regarded these data as missing at random or missing completely at random. Our reference sample included 233 qualified participants. SAS procedure Proc MI (SAS 9.2 added; SAS Institute, Cary, NC) was used to perform the multiple-imputation method. We used the default Markov chain Monte Carlo method, because the missing data pattern was arbitrary.

We included 44 of 485 available variables in the imputation model, paying attention to those that might be used in subsequent analyses of the imputed data sets.35 We included the dependent variable in the imputation model to reduce any artificial inflation of the relationship between the independent and dependent variables. The imputed values of the dependent variable did not provide additional error.37 Therefore, we calculated our primary outcomes by using the 3 key variables (quit status, CO reading, and days since the last cigarette) with their imputed values for observed self-classified and 7-day prevalence quit rates from an average of 5 imputation series, which allowed for an SD 5% wider than if we had used an infinite number of imputations.

We applied standard analytic techniques to our final observed and imputed data sets. A relative risk (RR) of >1 means that continued smoking is less likely to occur in the N-O-T or N-O-T+FIT group. We used a 2-proportion z test to determine statistical differences in quit rates between groups as follows: z = (p1 − p2)/SE, where SE = √[p1(1 − p1)/n1 + p2(1 − p2)/n2], p1 is the sample proportion from population 1, p2 is the sample proportion from population 2, n1 is the size of sample 1, and n2 is the size of sample 2.

RESULTS

Youths smoked approximately one-half of a pack of cigarettes per day (mean: 10.13 cigarettes per day; SD: 9.18 cigarettes per day) on weekdays and almost 1 pack per day (mean: 14.47 cigarettes per day; SD: 11.29 cigarettes per day) on weekends at program enrollment, which was consistent with our efforts to recruit daily teen smokers; 96% of youths smoked daily. We found moderate/high nicotine dependence scores. The mean age of the participants was 16.53 years (SD: 1.34 years) (Table 3).

Because of limited time in the funding period and the potential for cohort effects over an extended recruitment period, the final number of schools per condition was not ideal. Given the small number of schools for each treatment condition (BI: N = 6; N-O-T: N = 6; N-O-T+FIT: N = 7), we were limited in our approach to controlling for the clustering in the design. The majority of approaches to control for clustering, such as multilevel modeling and generalized estimating equations, require a larger number of units per condition for adequate power. Therefore, we did not use those analytic approaches. In an attempt to determine whether outcomes might have been attributable to school-level differences rather than intervention effects, we conducted a series of analyses of variance to examine whether there were significant baseline differences in 4 key variables (age, age at first cigarette use, nicotine dependence scores, and numbers of cigarettes per day) between schools, both within each treatment condition and between conditions. Within-group results suggested that 1 school in the BI group had a significantly lower average age, which was the result of having only 2 enrolled individuals, both 15 years of age (F5,57 = 7.26; P < .01); therefore, the school's average age was 15 years, with an SD of 0 years. Similarly, 1 N-O-T+FIT school had 3 enrolled individuals with an average age of 15.33 years (SD: 0.58 years), and 1 N-O-T+FIT school had 3 enrolled individuals with an average age of 15 years (SD: 1.00 years; F6,73 = 5.64; P < .01). There were no significant between-group differences (F2,16 = 0.81; P = .46). With respect to the age at first cigarette use, there were no significant differences either within groups (BI: F5,55 = 1.75; P = .14; NOT: F5,82 = 1.68; P = .15; N-O-T+FIT: F6,73 = 2.01; P = .08) or between schools (F2,16 = 1.17; P = .34). With respect to within-group levels of nicotine dependence, 1 N-O-T+FIT school had nicotine dependence scores significantly higher than average (mean: 7.67; SD: 0.58), and 1 school had nicotine dependence scores lower than average (mean: 4.00; SD: 2.52; within-group F6,69 = 3.04; P = .01). These differences in the N-O-T+FIT schools were attributable to 1 school having no female participants (the school with higher-than-average nicotine dependence scores) and the other school having no male participants (the school with lower-than-average nicotine dependence scores). There were no significant between-group differences in levels of nicotine dependence (F2,16 = 2.62; P = .10). Finally, 1 BI school had a lower-than-average value for the number of cigarettes per day (F5,55 = 2.60; P = .04), which was the result of the school enrolling 2 individuals, 1 who smoked an average of 4.3 cigarettes per day and 1 who smoked an average of 2.3 cigarettes per day. There were no other within-group differences in the numbers of cigarettes per day (N-O-T: F5,81 = 1.13; P = .35; N-O-T+FIT: F6,73 = 1.52; P = .19). There were no between-group differences in the numbers of cigarettes per day (F2,16 = 1.98; P = .17). Overall, these results suggested that the schools were largely equivalent with respect to key predictor variables, both within and between conditions, which increased our confidence that the results were not likely attributable to school-level effects.

Visual inspection of the data revealed an unusual distribution of participants in the precontemplation stage of change, particularly in the BI group (Table 3). To explore this issue, we conducted a series of t tests to determine whether the differences in the stages of change between groups were statistically significant. Results demonstrated that the BI group had significantly lower scores for stage of change, compared with the N-O-T group (t151 = −3.72; P < .001) and the N-O-T+FIT group (t141 = −4.17; P < .001). There was no significant difference between the N-O-T and N-O-T+FIT groups with respect to stage of change (t168 = −0.37; P = .71). Two-proportion z scores confirmed that there were significantly more participants in the precontemplation stage in the BI group, compared with the N-O-T group (z = −5.5; P < .001; RR: 2.49) and the N-O-T+FIT group (z = −6.29; P < .001; RR: 2.50). There were no significant differences in the numbers of participants in the precontemplation stage in the N-O-T and N-O-T+FIT groups. Further examination of the data showed that 1 BI school had a large number of boys in the precontemplation stage. To explore the potential effects of this confounding, we examined whether the large number of participants in the precontemplation stage in the BI group influenced quit outcomes in the BI group versus the other groups. Findings revealed that there were no significant differences in cessation outcomes between participants in the BI and N-O-T groups in the precontemplation stage (z = 0.94; P = .17) or between participants in the BI and N-O-T+FIT groups in the precontemplation stage (z = 1.12; P = .13).

In support of the primary study hypothesis (Table 4), the observed CO-validated, self-classified and 7-day point prevalence quit rates were consistent with our predicted trends 3 months after baseline. The 7-day quit rates were 4.76% for the BI group, 11.11% for the N-O-T group, and 13.75% for the N-O-T+FIT group. Imputed values for the BI, N-O-T, and N-O-T+FIT groups were 12.7%, 20.00%, and 33.75%, respectively. Overall, as predicted, the addition of the FIT adjunct decreased the risk of continued smoking twofold. Moreover, the RR values were >1 for both N-O-T and N-O-T+FIT groups, compared with the BI group, which indicated that youths in the BI group had a higher likelihood or risk of continuing to smoke after intervention than did youths in the N-O-T group (Table 5). The self-classified quit rates at 6 months after baseline were 15.87% for the BI group, 21.11% for the N-O-T group, and 31.25% for the N-O-T+FIT group. Although all groups showed increased proportions of abstainers, youths in the N-O-T+FIT group maintained a higher likelihood of quitting, compared with the risks of continued smoking in the BI group (RR: 1.97). Consistent with our hypothesis, youths in the N-O-T+FIT group had a greater likelihood of cessation, compared with those in the N-O-T group (RR: 1.48). Although the differences between the N-O-T and BI conditions were not statistically significant at 6 months, there was a large effect in favor of N-O-T (RR: 1.33) (Figs 2 and 3).

View this table:
TABLE 4

Quit Rates

View this table:
TABLE 5

RRs Overall and According to Gender

FIGURE 2
FIGURE 2

Observed self-classified point prevalence rates 3 months after baseline. All values are percentages.

FIGURE 3
FIGURE 3

Observed 7-day point prevalence rates 3 months after baseline. All values are percentages.

With consideration of all quit rate calculations at the 3-month follow-up assessment, there were marginally significant differences between boys in the N-O-T and BI groups. When we examined the N-O-T+FIT component, however, the added effect for boys was striking. For example, 7.89% of boys in the N-O-T group, compared with 23.68% of boys in the N-O-T+FIT group, reported 7-day point prevalence cessation (RR: 3.00). Imputed 7-day point prevalence values were 18.42% for the N-O-T group and 47.37% for the N-O-T+FIT group (RR: 2.71). In contrast, the corresponding observed 7-day quit rates for girls were 13.46% for the N-O-T group and 4.76% for the N-O-T+FIT group; the imputed values were 21.15% for the N-O-T group and 19.05% for the N-O-T+FIT group. Both calculations revealed RR values of <1.0, which suggests that the addition of physical activity to N-O-T did not decrease the risk of continued smoking for girls.

Across all measurement points and types of analyses, girls in the N-O-T group reported significantly higher quit rates than did girls in the BI group. The 7-day observed and imputed values for N-O-T and BI comparisons revealed remarkable RRs of ≥∞. The observed 7-day values showed that >13 times more girls in the N-O-T group, compared with girls in the BI group, quit smoking. Data strongly support the efficacy of N-O-T for smoking cessation among girls, with or without the physical activity supplement. However, the addition of more structured physical activity to N-O-T did not boost female cessation rates at the 3-month follow-up assessment. Similar trends were observed at the 6-month follow-up assessment. Depending on which quit rate calculation was used, it might be contended that the FIT component actually moderated quitting among girls.

DISCUSSION

Beyond examination of the influence of physical activity on cessation, this study served as an opportunity to examine quit rates under numerous conditions, including analyses similar to those used by Joffe et al16 in their randomized controlled trial of N-O-T. With the use of conservative nonimputed techniques, both N-O-T and N-O-T+FIT conditions showed consistently higher overall rates of biochemically validated cessation than did BI. The overall risk of continued smoking among youths in the BI group was almost 3 times higher. When we aggregated data across N-O-T conditions (N-O-T and N-O-T+FIT) and compared those data with data for the BI group, the observed 7-day quit rate for the aggregated N-O-T group was significantly higher than that for the BI group (BI:4.76%; N-O-T/N-O-T+FIT combined: 12.35%; z = 2.061; P = .020; RR: 2.59). The imputed 7-day point prevalence values showed similar results (BI: 12.70%; N-O-T/N-O-T+FIT combined: 26.47%; z = 2.555; P = .005; RR: 2.08).

When we analyzed data according to gender, we discovered that the N-O-T+FIT component decreased the risk of continued smoking fourfold among boys. This finding is consistent with evidence showing that physical activity is protective against smoking initiation and increased smoking levels among youths.10,–,12 In contrast to adult studies, these data provide further justification for enhancing physical activity as part of the N-O-T program for male participants. Our findings are consistent with other independent N-O-T investigations at comparable time points. A similar analysis by Joffe et al16 revealed statistically significant treatment effects of the N-O-T program on self-reported quit rates at the end of the program (RR: 1.26 [95% confidence interval CI]: 1.10–1.43), at 1 month (RR: 2.07 [95% CI: 1.68–2.56]), and at 12 months (RR: 1.58 [95% CI: 1.22–2.04]). A N-O-T study by Kohler et al38 in Alabama demonstrated a positive effect of the N-O-T program on 30-day, self-reported, quit rates at the end of the program, by using a conservative intent-to-treat analysis; youths in the N-O-T program had 4.2 times the odds of quitting, compared with control subjects. It is important to point out that the variations in quit rates were contingent on the criteria used (eg, self-classified versus 7-day and observed versus imputed). Although different classifications yielded slightly different rates, the trends were consistent. Equivalent to the ≥24-hour rates reported in previous N-O-T studies,18,19 the self-classified point prevalence rates were slightly higher than the 7-day rates. Point prevalence captures whether a participant is smoking at a particular point in time, and results can be validated biochemically, as in our study.32 The 7-day rates do not allow for slips, which often accompany cessation behavior. Moreover, although we measured CO levels, it is generally difficult to validate smoking behavior biochemically beyond 24 hours. Among the biochemical measurements to confirm smoking status, cotinine measurements, although they are expensive, are preferred by researchers because of cotinine's 17-hour half-life. CO has a shorter half-life of 1 to 4 hours.39 The various methods used to report quit rates cause some confusion regarding which reflect program outcomes most accurately. Point prevalence rates, such as the self-classified rates reported here, reflect real-time estimates of quit status and should be given consideration as a type of assessment of program impact. As presented by Hughes et al,40 we contend that multiple measures of outcomes provide a more accurate picture of effect size and that the effect size (and not simply the abstinence measure) should be a “major criterion for success” and clinical relevance. Because we found baseline differences in stages of change (ie, readiness), we examined cessation rates according to stage within each condition. We found a higher-than-expected frequency of subjects in the precontemplation stage in the BI group, especially among boys. We are not certain how to interpret these findings; it is plausible that in 1 BI school there was a “piggyback” effect with boys. These differences, however, provided us with an opportunity to explore outcomes according to stage. As reported by Cahill et al,41 the added value of adapting interventions to a smoker's stage of change is tentative. In their recent review, those authors reported that the evidence does not support the restriction of interventions only to smokers perceived to be in the preparation and action stages. Our study, in fact, found that both the BI and N-O-T conditions led to cessation across stages of change. In a N-O-T study by Dino et al,42 participants in the BI group in the preparation stage were 25 times more likely to have quit smoking at postbaseline assessments than were participants in the contemplation or precontemplation stages. In contrast, N-O-T was effective for youths regardless of baseline stage. Although our present study used a randomized design, the BI sample had a greater proportion of participants in the precontemplation stage, particularly in 1 school. The teens in the BI group who quit smoking represented all stages of readiness equally. Logically, N-O-T was more effective than BI for those in the preparation stage and beyond.

This study included only West Virginia youths, which might limit the generalizability of the findings. For instance, opportunities for physical activity in this rural mountainous state may be more limited than in other geographic areas. Moreover, West Virginia has among the highest US teen smoking rates and lowest physical activity rates.43 We must replicate the study with diverse groups of teens. Future studies also should give attention to co-occurring external influences on the intervention within the schools or communities. It is possible that the physical activity adjunct, rather than physical activity itself, was more culturally suitable for boys than for girls. The stage of change baseline difference also is a consideration for overall treatment effects, biased toward unusually high cessation outcomes for the boys in the BI group. Other unmeasured factors might be at play. Although we made great efforts to recruit and to randomize carefully, our sample size was small, as reflected in small numbers of participants per group and per condition. It has been suggested that designs such as that used in our study would benefit from addressing of intraclass correlations between randomized groups through analyses such as generalized estimating equations.44 When the numbers of randomized groups (schools in the present study) are small (eg, <20), however, generalized estimating equations and similar techniques are inappropriate.44,45 In such circumstances, generalized estimating equations are known to result in biased estimates of effects.46,47 The availability of intraclass correlation estimates for smoking outcome variables in school-based smoking interventions is limited. However, a few studies showed that school-level intraclass correlations for smoking and other health behaviors ranged from 0 to 0.075.48,–,50 Intraclass correlation coefficients in relevant school-based studies often are small,51 which enhances our confidence in our treatment effects.

CONCLUSIONS

We conclude that the N-O-T+FIT component had a large effect for boys, decreasing the risk of continued smoking almost fourfold by the end of the program. We also found significant overall effects of N-O-T, compared with BI, on quit status. With consideration of both the imputed and observed data, the overall, CO-validated, 7-day quit rates were between 13.75% and 33.75% for N-O-T+FIT, between 11.11% and 19.78% for N-O-T, and between 4.76% and 13.33% for BI. The effect sizes were large, in favor of the intervention conditions. Overall, girls quit more successfully with N-O-T, compared with BI (RR: >∞), 3 months after baseline, and boys responded better to N-O-T+FIT than they did to BI (RR: 2–3) or to N-O-T (RR: 1–2). In this study, the ≤5 minutes per session of additional effort by N-O-T facilitators to bolster physical activity had a significant program impact. A forthcoming article will explore whether the FIT component enhanced physical activity for N-O-T participants, paying particular attention to gender.

ACKNOWLEDGMENT

The Centers for Disease Control and Prevention provided funding for this study (grant U48/CCU310821).

Footnotes

    • Accepted June 8, 2011.
  • Address correspondence to Kimberly Horn, EdD, Department of Community Medicine, West Virginia University School of Medicine, 1 Medical Center Dr, PO Box 9190, Morgantown, WV 26505. E-mail: khorn{at}hsc.wvu.edu

  • This trial has been registered at www.clinicaltrials.gov (identifier NCT01242657).

  • FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

  • BI —
    brief intervention
    CO —
    carbon monoxide
    N-O-T —
    Not on Tobacco
    RR —
    relative risk
    CI —
    confidence interval
    FIT —
    Physical Activity Module

REFERENCES

  1. 1.↵
    1. Grimshaw GM,
    2. Stanton A
    . Tobacco cessation interventions for young people. Cochrane Database Syst Rev. 2006;(4):CD003289
  2. 2.↵
    1. Ussher M,
    2. Nunziata P,
    3. Cropley M,
    4. West R
    . Effect of a short bout of exercise on tobacco withdrawal symptoms and desire to smoke. Psychopharmacology (Berl). 2001;158(1):66–72
    OpenUrlCrossRefPubMed
  3. 3.↵
    1. Ussher MH,
    2. Taylor A,
    3. Faulkner G
    . Exercise interventions for smoking cessation. Cochrane Database Syst Rev. 2008;(4):CD002295
  4. 4.↵
    1. Marcus BH,
    2. Albrecht AE,
    3. King TK,
    4. et al
    . The efficacy of exercise as an aid for smoking cessation in women: a randomized controlled trial. Arch Intern Med. 1999;159(11):1229–1234
    OpenUrlCrossRefPubMed
  5. 5.↵
    1. Hughes JR
    . Combining behavioral therapy and pharmacotherapy for smoking cessation: an update. NIDA Res Monogr. 1995;150:92–109
    OpenUrlPubMed
  6. 6.↵
    1. Taylor AH,
    2. Ussher MH,
    3. Faulkner G
    . The acute effects of exercise on cigarette cravings, withdrawal symptoms, affect and smoking behaviour: a systematic review. Addiction. 2007;102(4):534–543
    OpenUrlCrossRefPubMed
  7. 7.↵
    1. Taylor A,
    2. Katomeri M
    . Walking reduces cue-elicited cigarette cravings and withdrawal symptoms, and delays ad libitum smoking. Nicotine Tob Res. 2007;9(11):1183–1190
    OpenUrlAbstract/FREE Full Text
  8. 8.↵
    1. Janse Van Rensburg K,
    2. Taylor A,
    3. Hodgson T,
    4. Benattayallah A
    . Acute exercise modulates cigarette cravings and brain activation in response to smoking-related images: an fMRI study. Psychopharmacology (Berl). 2009;203(3):589–598
    OpenUrlCrossRef
  9. 9.↵
    1. Prochaska JJ,
    2. Hall SM,
    3. Humfleet G,
    4. et al
    . Physical activity as a strategy for maintaining tobacco abstinence: a randomized trial. Prev Med. 2008;47(2):215–220
    OpenUrlCrossRefPubMed
  10. 10.↵
    1. Escobedo LG,
    2. Marcus SE,
    3. Holtzman D,
    4. Giovino GA
    . Sports participation, age at smoking initiation, and the risk of smoking among US high school students. JAMA. 1993;269(11):1391–1395
    OpenUrlCrossRefPubMed
  11. 11.↵
    1. Audrain-McGovern J,
    2. Rodriguez D,
    3. Moss HB
    . Smoking progression and physical activity. Cancer Epidemiol Biomarkers Prev. 2003;12(11):1121–1129
    OpenUrlAbstract/FREE Full Text
  12. 12.↵
    1. Larson NI,
    2. Story M,
    3. Perry CL,
    4. Neumark-Sztainer D,
    5. Hannan PJ
    . Are diet and physical activity patterns related to cigarette smoking in adolescents? Findings from Project EAT. Prev Chronic Dis. 2007;4(3):A51
    OpenUrlPubMed
  13. 13.↵
    1. Rodriguez D,
    2. Duton GF,
    3. Tscherne J,
    4. Sass J
    . Physical activity and adolescent smoking: a moderated mediator model. Ment Health Phys Act. 2008;1(1):17–25
    OpenUrlCrossRef
  14. 14.↵
    1. Brodersen NH,
    2. Steptoe A,
    3. Boniface DR,
    4. Wardle J
    . Trends in physical activity and sedentary behaviour in adolescence: ethnic and socioeconomic differences. Br J Sports Med. 2007;41(3):140–144
    OpenUrlAbstract/FREE Full Text
  15. 15.↵
    1. Dino G,
    2. Horn K,
    3. Goldcamp J,
    4. Fernandes A,
    5. Kalsekar I,
    6. Massey C
    . A 2-year efficacy study of Not On Tobacco in Florida: an overview of program successes in changing teen smoking behavior. Prev Med. 2001;33(6):600–605
    OpenUrlCrossRefPubMed
  16. 16.↵
    1. Joffe A,
    2. McNeely C,
    3. Colantuoni E,
    4. An MW,
    5. Wang W,
    6. Scharfstein D
    . Evaluation of school-based smoking-cessation interventions for self-described adolescent smokers. Pediatrics. 2009;124(2). Available at: www.pediatrics.org/cgi/content/full/124/2/e187
  17. 17.↵
    1. Dino G,
    2. Horn K,
    3. Abdulkadri A,
    4. Kalsekar I,
    5. Branstetter S
    . Cost-effectiveness analysis of the Not On Tobacco program for adolescent smoking cessation. Prev Sci. 2008;9(1):38–46
    OpenUrlCrossRefPubMed
  18. 18.↵
    1. Horn K,
    2. Dino G,
    3. Kalsekar I,
    4. Mody R
    . The impact of Not On Tobacco on teen smoking cessation: end-of-program evaluation results, 1998 to 2003. J Adolesc Res. 2005;20(6):640–661
    OpenUrlAbstract
  19. 19.↵
    1. Branstetter SA,
    2. Horn K,
    3. Dino G,
    4. Zhang J
    . Beyond quitting: predictors of teen smoking cessation, reduction and acceleration following a school-based intervention. Drug Alcohol Depend. 2009;99(1–3):160–168
    OpenUrlCrossRefPubMed
  20. 20.↵
    American Lung Association. N-O-T: Not On Tobacco. “Awards & Recognition.” Available at: www.notontobacco.com/award-and-recognition.php. Accessed May 9, 2010
  21. 21.↵
    1. Curry SJ,
    2. Emory S,
    3. Sporer AK,
    4. et al
    . A national survey of tobacco cessation programs for youths. Am J Public Health. 2007;97(1):171–177
    OpenUrlCrossRefPubMed
  22. 22.↵
    1. Eaton DK,
    2. Kann L,
    3. Kinchen S,
    4. et al
    . Youth risk behavior surveillance: United States, 2009. MMWR Surveill Summ. 2010;59(5):1–142
    OpenUrlPubMed
  23. 23.↵
    American Lung Association. N-O-T: Not On Tobacco: “About N-O-T.” Available at: http://www.notontobacco.com/about-n-o-t.php. Accessed May 10, 2011
  24. 24.↵
    1. Dino GA,
    2. Horn KA,
    3. Goldcamp J,
    4. Kemp-Rye L,
    5. Westrate S,
    6. Monaco K
    . Teen smoking cessation: making it work through school and community partnerships. J Public Health Manag Pract. 2001;7(2):71–80
    OpenUrlPubMed
  25. 25.↵
    1. Horn K,
    2. Dino GA,
    3. Gao X,
    4. Momani A
    . Feasibility evaluation of Not On Tobacco: the American Lung Association's new stop smoking program for adolescents. Health Educ. 1999;5(5):192–206
    OpenUrl
  26. 26.↵
    1. Dino G,
    2. Horn K,
    3. Goldcamp J,
    4. Maniar S,
    5. Massey C,
    6. Fernandez A
    . A state-wide demonstration of Not On Tobacco. J Sch Nurs. 2001;17(2):90–97
    OpenUrlAbstract/FREE Full Text
  27. 27.↵
    American Lung Association. Not On Tobacco (N-O-T): A Total Approach to Helping Teens Stop Smoking. New York, NY: American Lung Association; 2009
  28. 28.↵
    1. Horn K,
    2. Fernandes A,
    3. Dino G,
    4. Kalsekar I,
    5. Massey C
    . Adolescent nicotine dependence and smoking cessation outcomes. Addict Behav. 2003;28(4):769–776
    OpenUrlCrossRefPubMed
  29. 29.↵
    1. Horn K,
    2. Dino G,
    3. Fernades A,
    4. Kalsekar I
    . A study of Appalachian teen smokers who participated in the Not On Tobacco Program: 15-month post-baseline follow up on quit and reduction rates. Am J Public Health. 2004;94(2):181–184
    OpenUrlPubMed
  30. 30.↵
    1. Sussman S,
    2. Dent CW,
    3. Burton D,
    4. Stacy AW,
    5. Flay BR
    . Developing School-Based Tobacco Use Prevention and Cessation Programs. Thousand Oaks, CA: Sage; 1995
  31. 31.↵
    1. Velicer WF,
    2. Prochaska JO,
    3. Rossi JS,
    4. Snow MG
    . Assessing outcome in smoking cessation studies. Psychol Bull. 1992;111(1):23–41
    OpenUrlCrossRefPubMed
  32. 32.↵
    1. Velicer WF,
    2. Prochaska JO
    . A comparison of four self-report smoking cessation outcome measures. Addict Behav. 2004;29(1):51–60
    OpenUrlCrossRefPubMed
  33. 33.↵
    1. Clayton R,
    2. Ries Merikangas K,
    3. Abrams D
    . Introduction to tobacco, nicotine, and youth: the Tobacco Etiology Research Network. Drug Alcohol Depend. 2000;1(59 suppl 1):S1–S4
    OpenUrl
  34. 34.↵
    1. Gwaltney CJ,
    2. Bartolomei R,
    3. Colby SM,
    4. Kahler CW
    . Ecological momentary assessment of adolescent smoking cessation: a feasibility study. Nicotine Tob Res. 2008;10(7):1185–1190
    OpenUrlAbstract/FREE Full Text
  35. 35.↵
    1. Little RJ,
    2. Wang Y
    . Pattern-mixture models for multivariate incomplete data with covariates. Biometrics. 1996;52(1):98–111
    OpenUrlCrossRefPubMed
  36. 36.↵
    1. Rubin DB
    . Multiple imputation after 18+ years. J Am Stat Assoc. 1996;91(434):473–489
    OpenUrlCrossRef
  37. 37.↵
    1. von Hippel PT
    . Regression with missing Y's: an improved strategy for analyzing multiply imputed data. Sociol Methodol. 2007;37(1):83–117
    OpenUrlAbstract/FREE Full Text
  38. 38.↵
    1. Kohler CL,
    2. Schoenberger YM,
    3. Beasley TM,
    4. Phillips MM
    . Effectiveness evaluation of the N-O-T smoking cessation program for adolescents. Am J Health Behav. 2008;32(4):368–379
    OpenUrlPubMed
  39. 39.↵
    SRNT Subcommittee on Biochemical Verification. Biochemical verification of tobacco use and cessation. Nicotine Tob Res. 2002;4(2):149–159
    OpenUrlCrossRefPubMed
  40. 40.↵
    1. Hughes JR,
    2. Keely JP,
    3. Niaura RS,
    4. Ossip-Klein DJ,
    5. Richmond RL,
    6. Swan GE
    . Measures of abstinence in clinical trials: issues and recommendations. Nicotine Tob Res. 2003;5(1):13–25
    OpenUrlAbstract/FREE Full Text
  41. 41.↵
    1. Cahill K,
    2. Lancaster T,
    3. Green N
    . Stage-based interventions for smoking cessation. Cochrane Database Syst Rev. 2010;(11):CD004492
  42. 42.↵
    1. Dino G,
    2. Kamal K,
    3. Horn K,
    4. Kalsekar I,
    5. Fernandes A
    . Stage of change and smoking cessation outcomes among adolescents. Addict Behav. 2004;29(5):935–940
    OpenUrlCrossRefPubMed
  43. 43.↵
    Centers for Disease Control and Prevention. State Tobacco Activities Tracking and Evaluation System: behaviors, cigarette use (youth). Available at: http://apps.nccd.cdc.gov/statesystem/ComparisonReport/ComparisonReports.aspx?TopicID=100&MeasureID=121. Accessed May 10, 2011
  44. 44.↵
    1. Murray DM,
    2. Pals SL,
    3. Blitstein JL,
    4. Alfano CM,
    5. Lehman J
    . Design and analysis of group-randomized trials in cancer: a review of current practices. J Natl Cancer Inst. 2008;100(7):483–491
    OpenUrlAbstract/FREE Full Text
  45. 45.↵
    1. Bellamy SL,
    2. Gibberd R,
    3. Hancock L,
    4. et al
    . Analysis of dichotomous outcome data for community intervention studies. Stat Methods Med Res. 2000;9(2):135–159
    OpenUrlAbstract/FREE Full Text
  46. 46.↵
    1. Lu B,
    2. Preisser JS,
    3. Qaqish BF,
    4. Suchindran C,
    5. Bangdiwala SI,
    6. Wolfson M
    . A comparison of two bias-corrected covariance estimators for generalized estimating equations. Biometrics. 2007;63(3):935–941
    OpenUrlCrossRefPubMed
  47. 47.↵
    1. Fay MP,
    2. Graubard BI
    . Small-sample adjustments for Wald-type tests using sandwich estimators. Biometrics. 2001;57(4):1198–1206
    OpenUrlCrossRefPubMed
  48. 48.↵
    1. Murray DM,
    2. Stevens J,
    3. Hannan PJ,
    4. et al
    . School-level intraclass correlation for physical activity in sixth grade girls. Med Sci Sports Exerc. 2006;38(5):926–936
    OpenUrlCrossRefPubMed
  49. 49.↵
    1. Murray DM,
    2. Ronney BL,
    3. Hannan PJ,
    4. et al
    . Intraclass correlation among common measures of adolescent smoking: estimates, correlates, and applications in smoking prevention studies. Am J Epidemiol. 1994;140(11):1038–1050
    OpenUrlAbstract/FREE Full Text
  50. 50.↵
    1. Siddiqui O,
    2. Hedeker D,
    3. Flay BR,
    4. Hu FB
    . Intraclass correlation estimates in a school-based smoking prevention study: outcome and mediating variables, by sex and ethnicity. Am J Epidemiol. 1996;144(4):425–433
    OpenUrlAbstract/FREE Full Text
  51. 51.↵
    1. Lee J-H,
    2. Schell MJ,
    3. Roetzheim R
    . Analysis of group randomized trials with multiple binary endpoints and small number of groups. PLoS One. 2009;4(10):e7265
    OpenUrlCrossRefPubMed
Back to top

In this issue

Pediatrics
Vol. 128, Issue 4
1 Oct 2011
View this article with LENS
Email Article
Request Permissions
Article Alerts
Citation Tools
Effects of Physical Activity on Teen Smoking Cessation
Kimberly Horn, Geri Dino, Steven A. Branstetter, Jianjun Zhang, N Noerachmanto, Traci Jarrett, Melissa Taylor
Pediatrics Oct 2011, 128 (4) e801-e811; DOI: 10.1542/peds.2010-2599
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo
Print
Download PDF
Insight Alerts

Jump to section

Subjects