 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* Center for Health Research, Kaiser Permanente, Portland, Oregon
Oregon Research Institute, Eugene, Oregon
University of Rhode Island, Kingston, Rhode Island
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Methods. Both smoking and nonsmoking teens, 14 to 17 years of age, who were being seen for routine visits were eligible for this 2-arm controlled trial. Staff members approached teens in waiting rooms of 7 large pediatric and family practice departments within a group-practice health maintenance organization. Of 3747 teens invited at 
1 visits, 2526 (67%) consented and were randomized to tobacco intervention or brief dietary advice. The tobacco intervention was individually tailored on the basis of smoking status and stage of change. It included a 30-second clinician advice message, a 10-minute interactive computer program, a 5-minute motivational interview, and up to two 10-minute telephone or in-person booster sessions. The control intervention was a 5-minute motivational intervention to promote increased consumption of fruits and vegetables. Follow-up smoking status was assessed after 1 and 2 years.
Results. Abstinence rates after 2 years were significantly higher for the tobacco intervention arm, relative to the control group, in the combined sample of baseline smokers and nonsmokers (odds ratio [OR]: 1.23; 95% confidence interval [CI]: 1.03–1.47). Treatment effects were particularly strong among baseline self-described smokers (OR: 2.42; 95% CI: 1.40–4.16) but were not significant for baseline nonsmokers (OR: 1.25; 95% CI: 0.97–1.61) or for those who had "experimented" in the past month at baseline (OR: 0.95; 95% CI: 0.45–1.98).
Conclusions. Brief, computer-assisted, tobacco intervention during routine medical care increased the smoking cessation rate among self-described smokers but was less effective in preventing smoking onset.
Key Words: tobacco • adolescents • intervention • cessation • primary care
Abbreviations: OR, odds ratio • CI, confidence interval • GEE, generalized estimating equations • PTC, Pathways to Change
Tobacco use remains the number 1 cause of preventable morbidity and death in the United States, each year resulting in >430 000 deaths,1 >5 million years of potential life lost,2 and $50 billion of direct medical expenditures related to smoking alone.3 Among adults who have ever smoked daily, 82% first tried cigarettes and 53% smoked daily before age 18 years.4 According to recent estimates, almost one half of current adolescent smokers who continue to smoke regularly will die from a smoking-related disease.5
After several decades of decline, rates of smoking in the past month increased in the early 1990s for grades 9 through 12 and peaked in 1997 at 37% for grade 12. Smoking rates have since declined but remain unacceptably high at 
27% for 12th-graders.6 Rates of adolescent smoking are higher among whites (40%) than among either Hispanics (34%) or blacks (23%).5 Male and female adolescents have similar smoking rates, except among blacks (only 17% of high school girls smoke cigarettes, compared with 28% of boys).5 For most adolescent subgroups, rates of tobacco use must be at least halved to meet Healthy People 2010 targets.5
Tobacco use can be reduced among youths by preventing (or delaying) initiation and promoting cessation. The large volume of research on smoking prevention in schools and communities has been summarized in several government reports4,7 and meta-analyses.8 Although many interventions based on social influence or life skills training show significant short-term results, the effects dissipate over time. A comprehensive, long-term study by Peterson et al9 showed no effects for a multiyear, school-based, prevention program, which provides sobering evidence of the limitations of school-based programs alone. A recent smoking prevention intervention among families enrolled in 2 managed care organizations similarly found no effect on smoking susceptibility or experimentation among preteens.10 Comprehensive programs that include both school-based and community components show more durable effects.11,12 The clinical care setting is a potentially important, unexplored contributor to youth tobacco use prevention and cessation efforts.
In contrast to prevention programs, few adolescent cessation programs have been evaluated rigorously.13 School-based cessation programs, even when preceded by extensive formative research,14 typically recruit few smokers in that setting, experience high attrition rates, and produce low cessation rates.4 Very few studies have recruited adolescents in medical settings or used state-of-the-art approaches, such as computer-based expert systems.15 To date, these studies have been limited by nonexperimental designs and/or short-term outcomes.
The Teen Reach (Research Approaches to Cancer in a Health Maintenance Organization) program was a randomized, controlled trial of the long-term efficacy of brief clinician advice, the Pathways to Change (PTC) interactive computer program,15,16 and brief motivational counseling to reduce the prevalence of smoking among teens seen for routine medical care. This population-based, individually tailored intervention capitalized on the teachable moment present at the primary care physician visit and the attractiveness to teens of a multimedia, computer-based, interactive program. Other attributes of this strategy included (1) a focus on the whole population of adolescents (whether smoking at entry or not), (2) a stage-based, theory-driven intervention, (3) the ease and practicality of the intervention for clinicians, and (4) the relatively low cost of delivering the intervention to teens during routine medical care.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
450 000 individuals. Pediatrics and family practice departments at 7 Kaiser Permanente Northwest medical centers participated in the study during a 23-month recruitment period, from October 1997 through August 1999.
Participants
Figure 1 shows the flow of participants through the study. Study staff members scanned electronic appointment records to identify age-eligible teens with appointments at participating medical centers. Staff members invited adolescent primary care patients waiting to see their physician or allied health care provider to participate in a study of "healthy lifestyles and changes in health habits." Eligibility criteria included age 14 through 17 years, willingness to stay after the visit for 15 minutes, and no intention to leave the geographic area within 1 year. No incentives were offered. Subjects who agreed to participate signed consent forms and completed short, self-administered questionnaires. We assigned participants randomly to either a tobacco prevention/cessation intervention or a brief dietary intervention (both described below). Hollis et al17 provided additional details about recruitment and characteristics of the study sample.
|
Tobacco Intervention
We designed the tobacco intervention to reduce the prevalence of cigarette smoking by preventing smoking uptake among nonsmokers and increasing cessation among smokers. The intervention included 3 primary elements at the enrollment visit. (1) Staff members provided a written prompt (ie, a suggested 30–60-second advice message) to primary care clinicians to encourage teens to quit smoking or to not start. Clinicians were also asked to encourage the patient to talk briefly with a health counselor (study staff member) immediately after the visit. (2) Teens had a 10- to 12-minute session on a computer with the PTC expert system, which was developed by Prochaska, Velicer, and colleagues16,18,19 for use with adolescent nonsmokers and smokers. This multimedia interactive version of the print-based system has been successful in trials with adult smokers.20–23 The PTC program assessed the stage of readiness to begin smoking (for nonsmokers) or the stage of change to quit smoking (for smokers) and then delivered highly individualized advice and encouragement tailored to the teen's reported readiness stage. The program included brief (5–10-second), stage-relevant, teen testimonial movies and graphics designed specifically to appeal to teens. (3) Teens had 3 to 5 minutes of post-PTC motivational counseling with bachelor's or master's degree-level health counselors trained in motivational interviewing techniques.24 Handouts included a synopsis of the stage-relevant advice or encouragement generated by the PTC program, topic-specific information sheets generated by the computer program as requested by the participant, and small quit kits (eg, cinnamon sticks and Pez candy dispensers) to encourage nonsmoking.
In addition, we conducted 2 individual booster sessions with the PTC and health counselor during the next 11 months. Whenever possible, counselors met with teens at future visits within the intervention contact window to deliver the intended follow-up interventions. In practice, counselors conducted most follow-up intervention sessions via telephone, reading the text of the computer queries and advice messages to the teen and mailing the printed reports generated by the PTC program. Counselors made up to 25 contact attempts at a variety of day, evening, and weekend times, left up to 5 telephone messages, and sent up to 2 letters requesting callbacks.
Diet Intervention
Teens allocated to the diet intervention served as attention control subjects. Health counselors provided 3 to 5 minutes of motivational counseling to promote increased consumption of fruits and vegetables. The recommendations followed the National Cancer Institute Five-a-Day campaign. Counselors provided teens with 2 nutritional information brochures and a small snack-packet of fruit leather. This intervention did not include a computer component. We assumed that some teens would also receive brief tobacco advice as a part of usual care, but we did not prompt clinicians.
Intervention Training and Quality Assurance
Counselors used motivational interviewing techniques24 during the structured counseling sessions for both the tobacco and diet interventions. The lead health behavior interventionist, a master's degree-level counselor with substantial experience in adolescent health behavior changes, provided the initial counselor training and monitored quality and consistency over time. Interventionists conducted solo sessions only after completing training certification. In the field, interventionists documented components of intervention delivery, including whether they notified the clinician to deliver the brief advice message for the tobacco intervention group. These data were monitored to evaluate the quality of intervention delivery.
To assess the frequency of clinician tobacco cessation advice, research staff members asked participants in both study groups whether their clinician discussed any of the following topics during the visit: nutrition, sun exposure, tobacco use, and exercise. These questions were completed immediately after the consultation with the clinician and, for treatment subjects, just before receiving other intervention components (eg, the computer program).
Follow-up Assessments
A combination of mailed questionnaires and telephone interviews (for those who did not respond to the mailing) provided outcome data after follow-up periods of 1 and 2 years. One month before each annual anniversary of the teen's enrollment, research staff members mailed a questionnaire and a cover letter to the teen. Two weeks later, we mailed a second questionnaire if the first had not been returned. Two weeks after the second mailing, if necessary, study staff members attempted contact via telephone. Blinded study personnel conducted all follow-up assessments. To enroll more participants, we extended the recruitment period beyond our original plans and thus reduced the calendar time available for the 2-year assessment. The second annual follow-up period was therefore truncated by 2 to 3 weeks for the last 140 participants enrolled in the study. The mean follow-up period from study enrollment was 12.4 months (SD: 0.84) for the first assessment and 24.4 months (SD: 0.80) for the second assessment. Follow-up assessments were completed by the end of September 2001.
Measures
Outcomes
The primary outcome of the study was the proportion of teens who reported no smoking in the 30 days before the 1- and 2-year follow-up assessments. We hypothesized that abstinence rates would be higher in the tobacco treatment group, compared with the diet control group. Our projected sample size of 2538 assumed a usual care smoking prevalence at follow-up evaluations of 30% (actual value: 31.4%), 
of .05, 0.85 power, a 2-tailed test for the difference in proportions, and a 20% reduction in smoking prevalence in the treatment group. We also used self-reported 30-day use of cigarettes, cigars, pipes, and chewing tobacco at follow-up evaluations to assess abstinence from all forms of tobacco.
Covariates
Table 1 lists additional sociodemographic, tobacco use, and health-habit measures. We briefly summarize these measures here; Hollis et al17 described them in more detail. We based the depression screen on the 3 items recommended by Rost et al.25 Stage of acquisition for smoking and stage of smoking cessation were based on the algorithms described by Pallonen et al.26 Susceptibility to smoking in the next year was based on the algorithm described by Pierce et al.27,28
|
Analyses
We compared characteristics of the study groups at baseline by examining percentage distributions of categorical variables and means of continuous variables. We used a generalized estimating equations (GEE) approach for repeated categorical measures to assess the effects of the tobacco intervention across the 2 follow-up years. In these models, treatment group-time interactions provided the tests of hypothesized treatment effects. The primary test of the tobacco intervention's efficacy was the overall difference in rates of abstinence from smoking at follow-up evaluations. Secondary questions of interest, however, were whether the tobacco intervention reduced smoking onset among baseline nonsmokers, promoted cessation among baseline smokers, or both. Consequently, we examined GEE models for cigarette nonsmokers and smokers (30-day use) at baseline and for subcategories within these groups on the basis of a question about self-described smoking status. Most analyses were conducted with SAS software,31 although we used the Solas software package32 for the multiple-imputation procedure described below.
Missing Data Imputation
Because of the current debate regarding how to handle missing data in tobacco trials, we examined study outcomes by using 6 alternative methods for handling missing smoking outcome data.33,34 The first method used the available raw data (ie, "complete cases"), with no adjustments and no imputation for missing data. The second approach, propensity analysis, adjusted the intervention effect at each follow-up time for differences in participants' predicted propensity for being lost to follow-up monitoring. The propensity model was based on baseline age, gender, race/ethnicity, smoking status, stage of change, friends' smoking, smokers in the home, educational aspirations, BMI, exercise patterns, and depression. The third approach, the pattern-mixture method, adjusted the intervention effect for differences in the pattern of missing data (ie, no missing data, missing at either year 1 or year 2, or missing at both year 1 and year 2).
The last 3 methods imputed missing outcome data in various ways. Method 4 assigned the baseline value when follow-up smoking status was missing. Method 4 was equivalent to the traditional "intent-to-treat" procedure for baseline smokers, but baseline nonsmokers with missing outcomes were assumed to still be nonsmokers. Method 5, a single-imputation regression approach, used a combination of baseline variables (ie, same as for the propensity analysis) to predict and impute missing outcomes. Finally, a multiple-imputation regression approach used these same baseline predictor variables but created and then averaged 5 imputed values for each missing case. Multiple imputation provides a more accurate estimate of the variance for the intervention effect than does single imputation. Conclusions were largely consistent among the various missing-data procedures, and we there-fore relied on the multiple-imputation procedure for our primary presentation of the main outcomes.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
1 times in medical office waiting rooms consented to the study and were randomized. Table 1 presents baseline characteristics, smoking influences among family and friends, status on a depression screen, smoking status, stage of readiness for smoking acquisition or cessation, and smoking susceptibility. The tobacco and diet study arms were similar with respect to all items, with the exception of a small but significant difference in the proportions of subjects who screened positive for depressive symptoms.
Intervention Process Measures
Table 2 shows that a high percentage of clinicians in the tobacco arm received a written prompt to deliver brief tobacco advice. Immediately after the clinical encounter, patients in the tobacco arm were somewhat more likely than those in the diet arm to report that their clinician discussed the tobacco issue, although only 41% recalled such advice. As expected, reports of discussions of other lifestyle topics were similar in the 2 treatment groups.
|
Loss to Follow-up Monitoring
Both annual assessments achieved high response rates, with 2367 of 2526 subjects (93.7%) completing the 1-year follow-up assessment (963 by mail and 1404 by telephone) and 2218 of 2526 (87.8%) completing the 2-year follow-up assessment (743 by mail and 1474 by telephone). Figure 1 shows the reasons for nonresponse at each assessment. Loss to follow-up monitoring at 2 years was significantly greater (P < .001) in the tobacco arm (181 of 1254 subjects, 14.3%), compared with the diet arm (128 of 1272 subjects, 10.1%), primarily because those in the tobacco arm had more contact with staff members and thus more opportunities to decline participation.
Diet Treatment Outcomes
Self-reported servings of fruits and vegetables combined did not differ for subjects in the diet control group and the tobacco arms at baseline (3.66 vs 3.63, P > .78), 1 year (3.59 vs 3.62, P > .77), or 2 years (3.61 vs 3.60, P > .92). Similarly, no treatment effects were seen for intake of fruits alone at 1 year (0.99 vs 1.02, P > .40) or 2 years (0.97 vs 0.95, P > .80) or intake of vegetables alone at 1 year (1.50 vs 1.49, P > .92) or 2 years (1.50 vs 1.46, P > .48).
Tobacco Treatment Outcomes
Table 3 shows the proportions of participants who were smoke-free for 30 days at the 1- and 2-year assessments, with planned contrasts from a GEE analysis for treatment effects at each follow-up time, relative to baseline. We imputed missing values by using a baseline regression model with multiple imputations. The tobacco intervention, relative to the diet arm, increased significantly the proportion of all participants (ie, both smokers and nonsmokers at baseline) who were smoke-free for 30 days at the 1-year assessment. This modest treatment effect diminished somewhat by the time of the 2-year assessment but remained statistically significant. Treatment effects did not vary according to (ie, interact with) facility. A similar pattern of results applied when the outcome was defined as no tobacco use (including pipes, cigars, and smokeless tobacco) (data not shown).
|
1 cigarettes in the past 30 days at baseline (ie, experimenters, smokers, and recent quitters), the tobacco intervention produced significant effects at both the 1- and 2-year assessments, although the intervention had no effect on the small subgroup (n = 140) of self-described experimenters at baseline. In contrast, the tobacco intervention had a strong cessation effect among those who considered themselves to be smokers at baseline (odds ratio [OR]: 2.4; 95% confidence interval [CI]: 1.40–4.16). Conclusions were similar when the outcome was defined as no tobacco in the past 30 days. In separate posthoc analyses of baseline self-described smokers, the tobacco intervention produced a statistically significant effect among nonwhites (OR: 4.10; 95% CI: 1.01–16.71), despite the small sample size and large CI. The OR was nearly double that seen for whites (OR: 2.16; 95% CI: 1.14–4.08), although the CIs for whites and nonwhites overlapped. In process analyses for all participants in the tobacco treatment arm only, quit rates were higher among those who completed 1 (OR: 2.65; 95% CI: 1.89–3.73) or 2 (OR: 4.03; 95% CI: 2.87–5.65) booster calls, compared with those who completed none. The same was true for baseline nonsmokers who received 1 (OR: 2.72; 95% CI: 1.35–5.47) or 2 (OR: 2.89; 95% CI: 1.48–5.60) booster calls but not for the smaller group of baseline regular smokers (OR: 2.00; 95% CI: 0.83–4.90; and OR: 1.88; 95% CI: 0.73–4.83; respectively). Table 4 shows the ORs and CIs for treatment effects at the 2-year follow-up assessment with 6 alternate methods for handling missing data.33,34 Results and conclusions were generally similar across methods. For baseline nonsmokers, the regression-based imputation yielded somewhat lower ORs than did imputing the baseline value (ie, assuming that baseline nonsmokers with missing data were still not smoking). For self-described smokers at baseline, the traditional intent-to-treat approach of imputing the baseline value (ie, imputing smoking for missing values) yielded smaller but still strongly significant differences in 30-day abstinence rates between the diet and tobacco arms at 2 years (OR: 1.86; 95% CI: 1.07–3.23).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Among the entire sample of baseline smokers and nonsmokers, the computer-assisted tobacco intervention produced a significant, albeit modest, reduction in self-reported smoking prevalence at both the 1- and 2-year follow-up assessments. Intervention effects for self-described smokers at baseline were considerably stronger, however, and these effects were largely maintained over 2 years of follow-up monitoring. The 2-year quit rate of 24% among self-defined smokers was high for a relatively brief intervention. This result was, however, within the 22% to 25% cessation rate range reported for 4 adult smoking-cessation studies that used a print-based version of the multimedia expert system intervention used in this study.20–23 If this approach was implemented broadly and achieved similarly high recruitment and quit rates, then the expected impact on cessation at the population level would be striking.
Quit rates were relatively high (ie, 21% at 2 years) among control subjects who reported at least 1 cigarette in the past 30 days at baseline. This subgroup, however, included self-described experimenters, of whom almost one half (48%) in the control group reported abstinence 2 years later. Among self-described baseline smokers in the control group, the abstinence rate was much lower (11%). Clearly, the smoking patterns of many experimenters are in flux, which might argue for longer (eg, 2-year) follow-up intervals than are common in studies of adults (eg, 6–12 months).
As has been true for school-based, smoking prevention programs,9 our intervention was less effective in preventing smoking onset among baseline nonsmokers. Treatment effects were significant, although modest, at 1 year and became nonsignificant at the 2-year follow-up assessment. Similarly, no effects were seen among the small group of self-described recent experimenters who did not view themselves as smokers at entry. Why might tobacco prevention programs have so little effect on preventing onset among nonsmokers and those who are early in the tobacco uptake process? One possibility is that those who are still in the experimentation or social smoking phase fail to appreciate just how addicting tobacco really is. Previously, we reported17 that most of these experimenters at baseline (76.3%) had no intention to smoke in the future. Indeed, many intend to smoke just at parties, to smoke only for a while, or to quit well before any serious health effects would befall them. Unfortunately, many do become addicted and remain long-term smokers. Increasing awareness about the actual frequency of addiction and long-term smoking may be useful.
We found no evidence that our control condition, a 3- to 5-minute motivational intervention, increased intake of fruits and vegetables. The tobacco intervention, however, was well received by patients. In addition to the high recruitment rate, we found that 98% of those randomized to the tobacco arm used the computer program and received brief counseling. Although clinicians were prompted to advise treatment participants but not control subjects, treatment participants were only modestly more likely than control subjects (41% vs 28%) to report that their clinician discussed tobacco during the visit. It is possible that teens under-reported clinicians' tobacco advice, although teens' reports were given immediately after the visit and just before use of the computer program. Most participants (84%) completed 1 follow-up support call, and nearly one half completed 2 calls after extensive efforts to reach them. Other posthoc correlational analyses showed that those in the treatment arm who completed booster calls were more likely to be abstinent. This finding could reflect a true "dose-response effect" or perhaps could indicate simply that new and continuing smokers were less willing to take the calls. Although some teens were difficult to reach, it should be recalled that they were not volunteers who came to us for help with smoking. They were both smoking and nonsmoking patients with a wide range of interest in the tobacco issue who received a very brief intervention as a part of a routine visit.
Although the overall follow-up rates at 1 year (94%) and 2 years (88%) were high, loss to follow-up monitoring was somewhat greater in the treatment arm, compared with the control arm (14% vs 10%). We applied a variety of recommended methods to account for or impute missing data.33,34 We found that the various methods had relatively little effect on the outcomes or conclusions about treatment effects. For baseline smokers, the traditional approach of imputing the baseline value (ie, assuming that smokers with missing values were still smoking) yielded a lower, although still clearly significant, OR. For nonsmokers at baseline, multiple imputation was more conservative than substituting the baseline value and assuming that they were still not smoking. An even more conservative approach would be to assume that all baseline nonsmokers with missing follow-up values had started smoking. This assumption is hard to justify, however, given that 83% to 86% of baseline nonsmokers with follow-up data were still abstinent. We view the multiple imputation method as the more reasonable and sophisticated approach for handling missing data for a population that includes both smokers and nonsmokers at baseline. It is reassuring that the various missing-data procedures had minimal impact on the overall study conclusions.
A limitation here is that our study sample was largely white (78%). Power was insufficient for specific race or ethnicity subgroup comparisons, but treatment effects for nonwhites were significant and at least as strong as those seen for whites. Another limitation is that our end points were based on self-reported smoking status, which might understate or bias actual smoking rates. Ideally, we would have confirmed biochemically the smoking status at both follow-up points, but current biochemical verification procedures are not well suited for detecting the light and/or occasional smoking patterns typical among adolescents. Because our aim was to conduct a real-life and highly externally valid study, we judged that limiting participation to only those adolescent patients willing to consent to come back for future face-to-face follow-up visits to provide saliva samples would not be feasible and would likely reduce substantially the participation rate and generalizability. The Society for Research on Nicotine and Tobacco Subcommittee on Biochemical Confirmation35 also concluded that, in large-population, low-intensity, intervention trials, biochemical confirmation is often neither feasible nor desirable. Although self-report may inflate quit rates, the committee determined that the magnitude of such inflation is small and "rarely is differential across intervention conditions."35 Verification is warranted for teens and/or adults where there is an incentive to deceive, such as in diversion programs or when outcomes are shared with parents or other authority figures. This study had minimal incentives to deceive, because all participants were assured that their data would not be shared with either their parents or their clinician. In this relatively low-intensity, low-demand setting, we expect that misreporting would be modest, similar across groups, and unlikely to have accounted for the large treatment effects seen 2 years after treatment. The baseline age-specific prevalence of reported smoking in this population was also similar to that seen in anonymous statewide surveys, which suggests that most patients were willing to report honestly. However, reporting bias cannot be ruled out completely as a possible explanation for the findings.
The clinical context of our study and the individualized attention from trained counselors might have contributed to the strong effects on quitting among the youth smokers. A European study36 used an earlier version of the same computer program in a classroom setting, with no adjunct counseling or follow-up support, and found no prevention or cessation effects. That study suffered from having nonequivalent groups at baseline and a relatively short follow-up period. A follow-up report found that lack of engagement in the intervention was a predictor of subsequent smoking.37 One possibility is that individual patients in a clinical setting might be more likely to engage in the intervention than students in a classroom, although we could not assess this directly.
Although our intervention protocol was designed as a prototype for a practical, office-based strategy, research staff members delivered all aspects of the intervention other than the clinician's brief advice. In real-world practice, office staff members would need to assess smoking status, introduce the teen to the computer program, provide brief motivational counseling, and complete follow-up support. Highly motivated clinicians and support staff members could reasonably offer a similar intervention, but these components would likely prove challenging to integrate into most practice settings. Realistically, health care delivery systems will likely need additional research evidence and a more easily managed approach before they invest significant resources. Future research should explore ways to take advantage of the power and opportunity of clinical encounters while minimizing the time and training demands on clinicians. For example, clinical staff members could provide the first 3 A's (ie, ask, advise, and assess interest) and then refer interested teens for centralized, multisession, proactive, telephone counseling, combined with an interactive website that teens could access conveniently and repeatedly from home. Although they are potentially more practical, the efficacy of more centralized approaches needs to be determined, and we are currently testing 1 such model. We recommend currently that clinicians follow the US Public Health Service recommendations38 for adolescents and routinely ask about smoking, offer brief advice, assess smokers' interest in quitting, and provide brief assistance and follow-up monitoring as appropriate.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
We thank the following contributors to this work: Unto Pallonen, PhD, for helping develop and implement the expert system intervention; Steven Berg-Smith, for developing the motivational counseling components of the tobacco and diet interventions; Christine Catlin, for assuming lead interventionist/recruiter responsibilities; Donna White, for lead follow-up interviewing; Xiuhai Yang, MA, for data analysis; Kristine Funk, MS, RD, for manuscript production; Jennifer Coury, MA, for editing; Center for Health Research recruitment and intervention staff members; Kaiser Permanente Northwest physicians and medical office staff members; and the volunteer teen participants who made this research possible.
|
FOOTNOTES |
|---|
Reprints requests to (J.F.H.) Center for Health Research, Kaiser Permanente, 3800 N Interstate Ave, Portland, OR 97227. E-mail: jack.hollis{at}kpchr.org
Conflict of interest: Dr Velicer has received royalties from the University of Rhode Island from the licensing of smoking cessation programs.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C. L. Backinger, C. N. Michaels, A. M. Jefferson, P. Fagan, A. L. Hurd, and R. Grana Factors associated with recruitment and retention of youth into smoking cessation intervention studies--a review of the literature Health Educ. Res., April 1, 2008; 23(2): 359 - 368. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Pbert, A. J. Flint, K. E. Fletcher, M. H. Young, S. Druker, and J. R. DiFranza Effect of a Pediatric Practice-Based Smoking Prevention and Cessation Intervention for Adolescents: A Randomized, Controlled Trial Pediatrics, April 1, 2008; 121(4): e738 - e747. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Robinson, K. M. Emmons, E. T. Moolchan, and J. S. Ostroff Developing Smoking Cessation Programs for Chronically Ill Teens: Lessons Learned from Research with Healthy Adolescent Smokers J. Pediatr. Psychol., March 1, 2008; 33(2): 133 - 144. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. P. Tercyak, M. T. Britto, K. M. Hanna, P. J. Hollen, and M. M. Hudson Prevention of Tobacco Use Among Medically At-risk Children and Adolescents: Clinical and Research Opportunities in the Interest of Public Health J. Pediatr. Psychol., March 1, 2008; 33(2): 119 - 132. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Bancej, J. O'Loughlin, R. W Platt, G. Paradis, and A. Gervais Smoking cessation attempts among adolescent smokers: a systematic review of prevalence studies Tob. Control, December 1, 2007; 16(6): e8 - e8. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Jason, S. B. Pokorny, M. Adams, Y. Hunt, P. Gadiraju, T. Morello, M. Schoeny, and C. Dinwiddie Youth Caught in Violation of Tobacco Purchase, Use, and Possession Laws: Education Versus Fines Behav Modif, November 1, 2007; 31(6): 713 - 731. [Abstract] [PDF]
|
|
|
|
 |
|
 A. Jull How to write a commentary an editor's perspective Evid. Based Nurs., October 1, 2007; 10(4): 100 - 103. [Full Text] [PDF]
|
|
|
|
|
|
M. A. Hellems, M. J. Gurka, and G. F. Hayden Statistical Literacy for Readers of Pediatrics: A Moving Target Pediatrics, June 1, 2007; 119(6): 1083 - 1088. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. N. Hylkema, P. J. Sterk, W. I. de Boer, and D. S. Postma Tobacco use in relation to COPD and asthma Eur. Respir. J., March 1, 2007; 29(3): 438 - 445. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. J. Norman, M. A. Adams, K. J. Calfas, J. Covin, J. F. Sallis, J. S. Rossi, C. A. Redding, J. Cella, and K. Patrick A Randomized Trial of a Multicomponent Intervention for Adolescent Sun Protection Behaviors Arch Pediatr Adolesc Med, February 1, 2007; 161(2): 146 - 152. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Pbert, K. E. Fletcher, A. J. Flint, M. H. Young, S. Druker, and J. DiFranza Smoking Prevention and Cessation Intervention Delivery by Pediatric Providers, as Assessed With Patient Exit Interviews Pediatrics, September 1, 2006; 118(3): e810 - e824. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Werch, J. L. Grenard, J. Burnett, J. A. Watkins, S. Ames, and E. Jobli Translation as a function of modality: the potential of brief interventions. Eval Health Prof, March 1, 2006; 29(1): 89 - 125. [Abstract] [PDF]
|
|
|
|
 |
|
 T. Lancaster Clinician advice, an interactive computer program, and motivational counselling increased smoking cessation in teens Evid. Based Med., October 1, 2005; 10(5): 144 - 144. [Full Text] [PDF]
|
|
|
|
|
|
D. Kingston Clinician advice, an interactive computer program, and motivational counselling during routine medical visits increased reported smoking abstinence among teens Evid. Based Nurs., October 1, 2005; 8(4): 105 - 105. [Full Text] [PDF]
|
|
Read all P3Rs
| |||||||||||||||||||||||||||||||
