OBJECTIVE. The goal was to evaluate the degree to which a smoking prevention and cessation intervention was delivered by providers to adolescents in the pediatric office setting.
METHODS. Eight pediatric clinics in central Massachusetts were assigned randomly to either a special intervention (brief pediatric provider-delivered intervention plus peer counseling) or the usual care condition. Subjects (n = 2710) were adolescents 13 to 17 years of age, both smokers (smoked in the past 30 days) and nonsmokers/former smokers. The degree to which smoking prevention and treatment interventions were delivered by providers was assessed through patient exit interviews with adolescents after their clinic visits; interviews assessed the occurrence of 10 possible intervention steps.
RESULTS. The percentage of providers engaging in the smoking interventions differed significantly between the special intervention and usual care conditions, according to adolescent reports in the patient exit interviews. For nonsmokers/former smokers, overall patient exit interview scores were 7.24 for the special intervention condition and 4.95 for the usual care condition. For current smokers, overall patient exit interview scores were 8.40 and 6.24 for the special intervention and usual care conditions, respectively. Intervention fidelity of special intervention providers was 72.2% and 84.0% for nonsmokers/former smokers and current smokers, respectively.
CONCLUSIONS. Pediatric providers who receive training and reminders to deliver a brief smoking prevention and cessation intervention to adolescents in the context of routine pediatric primary care practice can do so feasibly and with a high degree of fidelity to the intervention protocol.
- smoking prevention and cessation
- pediatric providers
- intervention delivery
- patient exit interview
Smoking is the leading preventable cause of illness and death in the United States.1,2 It is a serious pediatric health issue,3 because dependence begins during childhood or adolescence in the majority of cases.4 The most powerful predictor of adult smoking is smoking during adolescence.5 Nearly 90% of adult smokers smoked their first cigarette and 71% were daily smokers before 18 years of age6; 60% started before 14 years of age.7 The most recent National Youth Tobacco Survey found that 22% of high school students reported current cigarette smoking, defined as use in the past 30 days.8 In 2001, 19% of 12th-grade students reported daily smoking.9 Almost 75% of adolescent daily smokers who believe they will not be smoking in 5 years find that they are unable to quit.6 The American Academy of Pediatrics has issued a number of policy statements relevant to the problem of adolescent smoking and nicotine dependence in recent years, attesting to the fact that pediatricians as a group are well positioned to take an active role in addressing this issue.10–12
Although cessation is less common among adolescents than adults, interest in quitting is strong; nearly three fourths of adolescent smokers have thought seriously about quitting, 64% reported having made a quit attempt, and 40% of daily smokers reported having tried to quit at least once and failing.6 Among 17-year-old smokers, 40% reported that they would be interested in cessation treatment.6 Cessation and successful abstinence are more common among smokers who begin smoking at ≥17 years of age,13 which emphasizes the need to prevent smoking initiation during adolescence.
Primary care settings provide key opportunities to deliver smoking prevention and cessation interventions, because pediatricians, family practitioners, and other clinicians are in a unique position to identify and to intervene therapeutically with young tobacco users.14 Between 63% and 85% of adolescents are seen for preventive care visits each year.15–17 A study found that 63% of adolescents 14 to 17 years of age were seen in primary care settings over the course of 1 year and 83% over 2 years.18 Among those with ≥1 visit, 60% had ≥2 visits during the year. Because of their authority, credibility, and long-term relationships with patients, pediatric providers are afforded many clinical opportunities and teachable moments to deliver smoking prevention and cessation interventions during routine care.7
There is strong evidence that brief counseling by physicians, dentists, and other health care providers increases smoking cessation rates for adult patients,19–30 and evidence-based clinical practice guidelines regarding how to treat nicotine dependence among adults are available for clinicians and health care organizations.26 In contrast, there are sparse data regarding the efficacy of brief clinician interventions in treating tobacco use in adolescence. At present, expert opinions rather than empirical data are used to guide clinical interventions for young smokers. Existing guidelines recommend routinely that clinicians deliver strong messages regarding abstinence from tobacco and cessation of use to children and adolescents.7,13,26,31–39 Despite consensus among experts regarding these recommendations, many pediatricians and family practitioners do not assess and counsel youths about tobacco use routinely.40–48 Pediatricians report a greater tendency to encourage children and adolescents not to start smoking than to intervene with those who already smoke.40 Reasons cited for the low rates of addressing smoking among youths include lack of clear clinical practice guidelines, lack of training, perception of poor effectiveness of interventions, perception of low self-efficacy in delivering effective interventions, and lack of reimbursement or other incentives for delivering treatment.40
The objective of this article is to present the results of an evaluation of the degree to which a smoking prevention and cessation intervention was delivered by providers to adolescents in a pediatric office setting during routine medical care, based on adolescent reports in patient exit interviews (PEIs). These data were collected as part of a larger, randomized, controlled trial evaluating the impact of a smoking prevention and cessation intervention delivered by pediatric providers and peer counselors (local college students who were ex-smokers) to adolescents seen in pediatric practice. Sites were stratified according to size and assigned randomly to condition within strata. The study was approved by the University of Massachusetts Medical School human subjects institutional review board.
Setting, Intervention, and Training Program
Eight pediatric clinics located throughout central Massachusetts were randomized to 1 of 2 conditions, namely, (1) special intervention (SI), a provider-delivered smoking prevention and cessation counseling intervention tailored to the stage of smoking acquisition of each adolescent, with peer counseling focusing on the social aspects of smoking and development of the ability to resist social pressures to smoke; or (2) usual care (UC), the comparison condition. At the 4 intervention (SI) study sites, an office management system was established to support delivery of the intervention. All pediatric providers at all 8 study sites (both SI and UC), including pediatricians, nurse practitioners, physician's assistants, and pediatric residents, received a baseline survey, for collection of demographic data and evaluation of their attitudes toward and self-efficacy in intervening with adolescents regarding smoking, knowledge of smoking prevention and cessation in adolescents, and self-reported practice behaviors. Providers at the SI sites then received training in the brief provider-delivered smoking prevention and cessation counseling intervention. Providers for the UC condition did not receive any training.
The brief provider-delivered smoking prevention and cessation intervention was based on the 5A model recommended by the US Public Health Service clinical practice guideline26 and the American Academy of Pediatrics.10 It was adapted for use with adolescents by pediatric health care providers through focus groups with adolescents. The counseling approach used was patient-centered, which means that the provider was trained to engage the adolescent in a discussion about smoking that was nonthreatening, encouraged the adolescent to consider his or her beliefs and intentions regarding smoking, and allowed the adolescent to take in the provider's advice and to determine what course of action he or she would like to take. An intervention algorithm was developed with pediatric provider input and feedback that assessed the adolescent's smoking status and tailored the intervention on the basis of the individual's current smoking status and responses to key questions asked. Figure 1 presents the provider algorithm that describes the intervention steps to be taken by providers in the SI condition.
The pediatric provider training program consisted of a 1-hour group training tailored to the individual practice and held on-site in each pediatric office, followed in 2 to 4 weeks by an individual feedback session. The group training included a brief didactic presentation on the problem of smoking among adolescents, the role of health care providers, nicotine addiction, health risks related to smoking, health and social risks of relevance to teens, and benefits of not smoking/quitting reported by adolescent focus groups conducted as part of the randomized, controlled trial, an overview of the current US Public Health Service clinical practice guideline for treating tobacco use and dependence26 endorsed by the American Academy of Pediatrics,10 and an introduction to the peer counseling intervention. The intervention algorithm was reviewed in detail and a demonstration was given. The providers then broke into pairs to practice the intervention. Each provider attending the training received a 30-minute individual session with the project director 2 to 4 weeks after the group training, to practice the intervention and to receive feedback on intervention techniques. Practice and feedback continued until the clinician could complete the intervention protocol successfully. At the SI sites, the study algorithm was attached to the top of the study participants' medical record by the on-site research assistant; this did not occur at the UC sites. The study algorithm instructed SI providers to screen adolescents for smoking status and to provide a brief, patient-centered, smoking prevention/cessation intervention tailored to their stage of smoking acquisition and/or stage of change within the course of their visit.
Subjects and Data Collection
Eligible adolescents were patients 13 to 17 years of age (inclusive) at study entry, regardless of smoking status, who were scheduled for a visit at 1 of the 8 pediatric study sites. Smokers were defined as adolescents who reported having smoked at least a puff of a cigarette in the 30 days before baseline assessment. For eligible adolescents <16 years of age, a consent form was signed by both the parent and the teenager. For adolescents ≥16 years of age, the consent form required the teenager's signature only. All eligible adolescents who agreed to participate were informed that they may or may not be asked to breathe into a carbon monoxide monitor for biochemical validation of their smoking status and then completed a baseline survey before seeing their pediatric health care provider. After the visit with the pediatric health care provider, the adolescent was interviewed by the research assistant with the PEI, for assessment of the content and quality of the intervention received from the provider (both SI and UC). Adolescents completed follow-up assessments at 6 and 12 months. Measures used in the analyses for the present study included the PEI and selected items from participant and provider surveys.
The main data source for determining the degree to which smoking prevention and cessation counseling interventions occurred was patient report of intervention received, as assessed with PEIs after the provider (SI and UC conditions) visits (Fig 2). The PEI was administered by research assistants in person immediately after the patient's study entry clinic visit or by telephone within 48 hours after the visit. The PEI assessed whether specific intervention steps were used by the provider, tailored to the adolescent's smoking status (ie, nonsmoker/former smoker or current smoker), allowing for evaluation of the delivery of individual counseling steps and the development of a sum of steps out of a possible 10 points (the PEI index score).
Participants completed surveys at baseline and at 6- and 12-month follow-up times. The following variables were assessed. Demographic information included age (date of birth), gender, race/ethnicity, and school performance (baseline only). Confidence in ability to stop smoking (smokers) or remain smoke-free (nonsmokers) was assessed with a 10-point Likert scale ranging from 0 (not at all confident) to 10 (very confident), in response to the question, “How confident are you that you can remain a nonsmoker/could stop smoking totally and for good?” (L. Lamkin, MPA, CSADP, oral communication, 2000).
Depressive symptoms were assessed with the Children's Depression Inventory-Short Form (CDI-S) self-report measure.49 Each of the 10 items is scored from 0 to 2 (0 = absence of symptom, 1 = mild symptom, and 2 = definite symptom within the past 2 weeks). The CDI has been found to be reliable (Cronbach's α ranging from .71 to .89 and 1-week and 1-month test-retest reliabilities of >0.60) and to correlate well with other measures of childhood depression. Attention-deficit disorder (ADD)/attention-deficit/hyperactivity disorder (ADHD) was assessed with the question, “Have you ever been told by your pediatrician/nurse practitioner, parent, or school that you have ADD or ADHD?” Whether medication was taken was also assessed.
Current smoking status was determined with the question, “Which of the following best describes your experiences with cigarette smoking?” If adolescents responded either, “I smoke occasionally now” or “I smoke regularly now,” then they were classified as a current smoker. If they responded, “I have never smoked cigarettes, even 1 or 2 puffs,” or “I have smoked or tried cigarette smoking, even 1 or 2 puffs, before but not in the past year,” then they were classified as a nonsmoker. If they responded, “I have smoked or tried cigarettes in the past year but not in the past 30 days,” then they were classified as a former smoker.50,51
Smoking environment was assessed with 2 items. The first addressed smoking in the home, “Does anyone you live with smoke cigarettes?,” for which the participant checked as many of the following as was appropriate: no one I live with smokes; my mother, stepmother, or the woman guardian I live with smokes; my father, stepfather, or the man guardian I live with smokes; a brother or sister I live with smokes; someone else I live with smokes (L. Lamkin, MPA, CSADP, oral communication, 2000). The second item addressed peer smoking, “How many of your friends smoke cigarettes?,” with responses of none, only a few, less than half, about half, more than half, or almost all.52
Additional items for smokers included the following. Stage of change was assessed with the question, “Do you plan to quit smoking in the next…,” for which the participant selected 1 of the following 7 response categories: 6 months, 3 months, 30 days, 1 week, I am in the process of quitting, I have quit already, or I do not plan to quit in the next 6 months (adapted from the work by DiClemente et al53). Number of cigarettes smoked in the past 30 days was assessed with a combination of 2 questions. The first was, “During the past 30 days, on the days you smoked, how many cigarettes did you smoke per day?,” for which the participant selected one of the following responses: I did not smoke cigarettes during the past 30 days (0 cigarettes), <1 cigarette per day (0.5 cigarettes), 1 cigarette per day (1 cigarette), 2 to 5 cigarettes per day (3.5 cigarettes), 6 to 20 cigarettes per day (13 cigarettes), or >20 cigarettes (1 pack) per day (21 cigarettes).54 The midpoint was selected for coding purposes, to represent the range of responses, as indicated in parentheses above. The second question was, “During the past 30 days, on about how many days did you smoke cigarettes?,” for which the participant selected one of the following responses: 0 days (0 days), 1 or 2 days (0.5 days), 3 to 5 days (4 days), 6 to 9 days (7.5 days), 10 to 19 days (14.5 days), 20 to 29 days (24.5 days), or all 30 days (30 days).55 Again, the midpoint was selected for coding purposes, as indicated in parentheses.
The modified Fagerstrom Tolerance Questionnaire56 assesses the level of nicotine dependence among adolescents with 7 items, using a 5-point Likert scale for 6 items and a yes/no response regarding smoking during the first 2 hours of the day. A total score is obtained by summarizing raw scores with scoring codes designed to make the score of each individual item comparable to those of the original Fagerstrom Tolerance Questionnaire. Cutoff points for the total score, indicating the level of dependence, are as follows: 0 to 2, no dependence; 3 to 5, moderate dependence; 6 to 9, substantial dependence. Test-retest reliability was indicated by r = 0.71 (2-month interval)57 and internal consistency by Cronbach's α of .75.58
The Hooked on Nicotine Checklist (HONC)4,59 is a self-report measure of symptoms of nicotine dependence suggesting loss of autonomy over tobacco use, defined when either the physical or psychological sequelae of tobacco use present a barrier to quitting. A positive response to any of the 9 yes/no HONC items indicates a loss of autonomy and onset of dependence. Test-retest reliability has been found to be favorable (1-week κ = .61; 2-week κ = .75), and content and construct validity has been shown in screening for both adolescents and adults.60–62
Additional items for nonsmokers and former smokers included the following. Intention and efficacy expectations about future smoking, as assessed with a 3-item instrument, were used to predict which never-smokers are likely to start smoking, with the following questions. “Do you think you will try a cigarette soon?” “Do you think you will be smoking 1 year from now?” “If one of your best friends were to offer you a cigarette, would you smoke it?”63 The 4 response categories were definitely yes, probably yes, probably not, and definitely not. Nonsmokers who answered “no” to the first question and “definitely not” to the other 2 were considered committed never-smokers; all others were considered susceptible never-smokers. The measure has good validity (Cronbach's α = .84).63
All pediatric providers at the 8 participating sites (both SI and UC) received a survey at baseline and 12 months after the baseline survey. The survey assessed demographic variables (gender, age, and clinician type) (baseline only) and evaluated providers' attitudes toward and self-efficacy/confidence in intervening with adolescent smoking, knowledge regarding smoking prevention and cessation in adolescents, and self-reported practice behaviors.
Sites were compared with respect to the characteristics of providers and adolescent participants with either χ2 tests or Fisher's exact test, when the analysis contained small cells, for categorical data (eg, ethnicity) and t tests for dimensional data (eg, age). Multivariate analyses included backward stepwise linear regression for PEI total scores and backward stepwise logistic regression for individual PEI items. Regressions were adjusted for clustering according to site. To correct for multiple comparisons, Bonferroni corrections were used. Because the use of Bonferroni corrections based on all analyses conducted would suggest P values that were too conservative, we corrected within “families” of analyses. For example, for the 21 test results reported in Table 1, we used a Bonferroni correction that suggested a significance level of .05/21 = .002 for those tests. For the backward stepwise regressions, we corrected for the total number of items entered in step 1, regardless of the number of steps required to reach a final model. To assess the possibility that responses to the PEI could be attributed to variables in addition to, or instead of, the adolescent's study condition (SI versus UC), regression analyses were conducted with adjustment for clustering according to sites. Linear regression was used to analyze predictors of average PEI scores; logistic regression was used to analyze predictors of individual PEI question scores. The predictor variables considered in all analyses were gender, age, diagnosis of ADD/ADHD, CDI-S score, confidence in ability to stop smoking (for current smokers) or remain smoke-free (for nonsmokers and former smokers), grades in school, and study condition (SI or UC). For questions relevant to smokers and nonsmokers alike, smoking status (smoker versus nonsmoker) was included in the analysis. For questions relevant to current smokers only, the number of cigarettes smoked in the previous month, stage of change, HONC total score, and Fagerstrom Tolerance Questionnaire total score were also included.
Participant characteristics at baseline are presented in Table 1. A total of 2710 adolescents 13 to 17 years of age, regardless of smoking status, were recruited in the 8 pediatric offices. Baseline data were available for 2710 adolescents, and PEI data were available for 2707 adolescents. Participants in the UC and SI conditions were similar with respect to most variables but differed with respect to gender (more female subjects in the SI condition; P < .005), ethnicity (more Hispanic participants in the SI condition; P = .03), whether someone else in the household smoked (greater reported prevalence in the SI condition; P = .04), and ADD/ADHD (greater reported prevalence of ADD/ADHD in the SI condition; P = .008). None of these significance levels, however, was less than the Bonferroni-corrected cutoff level for Table 1 of .002. A total of 261 participants reported having smoked in the past 30 days.
Providers in the UC and SI conditions did not differ with respect to any characteristics assessed at baseline (Table 2). The majority of providers were women, and providers were 42 years of age, on average. There were 48 pediatricians, 10 nurse practitioners, 1 physician assistant, and 2 pediatric residents (data on clinician type were missing for 3 providers). Providers reported greater confidence overall in preventing adolescents from starting to smoke than in helping adolescents stop smoking.64
Ninety-eight percent of the PEIs were completed in person (n = 2655); 2% (n = 52) were completed by telephone after the visit. The PEI total scores were found to be distributed normally (skewness: −0.241; kurtosis: −0.397), with a mean of 6.41 (SD: 2.00) and a median of 7.00.
Fidelity to Intervention (SI Condition)
The mean PEI score for providers in the SI condition was 7.24 (SD: 1.55) of 10 intervention steps for nonsmokers/former smokers (72.4% of intervention) and 8.40 (SD: 1.93) of 10 intervention steps for current smokers (84.0% of intervention). The percentages of adolescents in the SI condition who reported that their provider engaged in each of the steps in the intervention protocol are reported in Table 3. Among current smokers, the most frequently reported steps taken were asking whether they were interested in quitting and stating that not smoking is the best thing they can do to protect their health; the least frequently reported was asking what the adolescent does not like about smoking. Among nonsmokers/former smokers, the most frequently reported steps taken were telling them it was great they were not smoking and stating that not smoking is the best thing they can do to protect their health; the least frequently reported was talking about nicotine addiction.
Comparison of Interventions Delivered by Providers in the SI and UC Conditions
The percentages of providers who engaged in the recommended smoking intervention steps differed significantly between the SI and UC conditions, according to adolescent reports on the PEI. For nonsmokers/former smokers, overall PEI scores were 7.24 (SD: 1.55) for the SI condition, compared with 4.95 (SD: 1.69) for the UC condition [t(1900) = 30.40; P < .001]. For current smokers, overall PEI scores were 8.40 (SD: 1.93) and 6.24 (SD: 2.12) for the SI and UC conditions, respectively [t(173) = 7.05; P < .001]. The percentages of adolescents in the SI and UC conditions who reported that their provider engaged in each of the intervention steps are presented in Table 3, along with a comparison across conditions. PEI data indicate that SI providers were more likely than UC providers to engage in the steps recommended to be taken with all adolescents, regardless of smoking status. For current smokers, all except 2 steps (asking how much the teen smokes and sharing additional reasons for not smoking) were performed more frequently by SI providers than by UC providers; for nonsmokers and former smokers, all steps were performed more frequently by SI providers, compared with UC providers.
In addition to reporting the practice behavior of providers, participants were asked about how well they thought their provider understood how they feel about smoking and the degree to which they believe it is their provider's job to discuss smoking with teens. Of participants in the SI condition, 66% reported feeling their provider understood their feelings about smoking very well, compared with 49% of participants in the UC condition [χ2(4) = 77.55; P < .0001]. SI participants were more likely to report that it was much or very much their provider's job to discuss smoking with teens, compared with UC participants [80% and 74%, respectively; χ2(4) = 15.47; P = .004]. In terms of time spent by the provider in discussing smoking during the visit, participants in the SI condition reported that their providers spent 4.3 minutes (SD: 3.13 minutes), compared with 2.9 minutes (SD: 2.85 minutes) in the UC condition (z for Mann-Whitney U = −13.66; P < .0001).
Predictors of PEI Score and Individual Intervention Steps
Predictors of PEI score and individual intervention steps are presented in Table 4. For the average PEI score, only being in the SI condition was associated with increased PEI averages. Greater depression scores on the CDI-S might have been associated with lower PEI averages but did not reach significance. Three questions were asked of all adolescents regardless of their smoking status. Being in the SI condition or being a smoker was associated with an increased likelihood of reporting that their provider discussed cigarette smoking, whereas greater symptoms of depression were associated with a decreased likelihood of this occurring. Providers were more likely to tell the adolescent that the discussion about smoking would be private when the teen was in the SI condition and female. Adolescents reported that they were more likely to be asked by the provider if they smoked or had ever tried smoking if they were in the SI condition.
Seven questions on the PEI were asked only of those adolescents who identified themselves to the provider as smokers (Table 4). Smokers' responses indicated that providers were more likely to ask how much the adolescent smoked if the adolescent endorsed more symptoms of nicotine dependence on the HONC, and both SI condition and increasing symptoms of nicotine addiction, as measured with the HONC, were associated with greater likelihood of providers asking smokers if they were interested in quitting. Except for asking how much they smoked and sharing additional reasons for not smoking, being in the SI condition was the only variable likely to predict that providers would engage in the recommended smoking interventions with self-reported smokers.
Being in the SI condition also increased the likelihood of providers conducting all 7 of the recommended intervention steps for nonsmokers and former smokers (Table 4). The only other significant predictor was increased depression scores, which was associated with decreased likelihood of the adolescents reporting that the provider congratulated them for not smoking and discussed addiction with respect to friends who smoke or in general.
Our findings suggest that a brief smoking prevention and cessation intervention can be implemented feasibly by providers for adolescents in routine pediatric primary care clinical practice, with a high degree of fidelity, as assessed with adolescent reports in PEIs. The majority of intervention steps were reported by adolescents in the SI condition to have been delivered by their providers, compared with fewer intervention steps reported by adolescents in the UC condition. Overall and for most individual steps, the reported frequency of intervention was significantly greater among SI participants, compared with UC participants, which indicates a strong effect of provider training and intervention implementation on the dose of smoking prevention and cessation intervention received by adolescents. In evaluation of the efficacy of an intervention, key issues are the extent to which the intervention components are implemented according to protocol and the dose of exposure to the intervention received by the targeted population. Such process information helps to explain why a program has succeeded or failed to meet its objectives. It is particularly important given the issues inherent in integrating an innovation into a busy, established, clinical setting. The findings of the present study suggest that the provider-delivered intervention was integrated effectively into busy, established, clinical practices and that adolescents in the study who were seen by these providers received a large dose of the intervention. Any impact of a smoking intervention on adolescent smoking can be expected only to the extent that the intervention is delivered as intended by providers and is received by study participants. Therefore, the high degree of intervention fidelity found for SI providers, both in absolute terms and in comparison with the UC condition, allows for a fair evaluation of the efficacy of the SI intervention in reducing smoking at 6- and 12-month follow-up assessments.65 In addition, it allows for an evaluation of the impact of the degree of intervention fidelity (ie, dose of intervention recalled by the adolescent) on smoking status.
The findings for the UC condition are of interest because they provide some insight into the usual practices of pediatric providers regarding the delivery of smoking prevention and cessation interventions to their adolescent patients. One previous study found that 77% of adolescents reported that their pediatric or family provider had discussed cigarettes/smoking,66 which is identical to the 77% of adolescents in the UC condition in the present study who reported being asked if they ever smoked or had tried smoking. However, no other study has assessed adolescent reports of specific smoking prevention and cessation intervention steps taken by pediatric providers during routine care. The results of this study begin to shed light on the degree to which providers engage in specific smoking prevention and cessation counseling steps with adolescents. It was found that pediatric providers in the UC condition engaged in each of the intervention steps assessed to varying degrees, albeit at much lower rates than in the SI condition. For smokers, the rates were notably low for asking adolescents what they like and do not like about smoking (ie, exploring ambivalence) and assessing their interest in quitting. For nonsmokers and former smokers, adolescents reported that providers infrequently asked their reasons for not smoking, shared additional reasons, or discussed nicotine addiction. The exception was UC providers asking smokers how much they smoke, giving smokers reasons for quitting, and congratulating nonsmokers and former smokers for not smoking, each of which occurred at relatively high rates in both the UC and SI conditions.
Patient PEI reports of smoking prevention and cessation intervention received from their pediatric provider in the UC condition in the present study indicate lower rates of intervention than those reported in surveys of pediatricians and family physicians,46,47 in which the majority of providers reported assessing smoking status for most or all of their patients, encouraging most or all not to start smoking, and providing advice to stop smoking for those who smoke. This finding is consistent with the adult tobacco literature, which also found that, in comparison with patient reports, surveys of physicians usually indicated higher levels of assessing smoking status and providing smoking cessation advice and counseling.67–71 It seems that a similar pattern of pediatric providers reporting greater intervention than patients is in effect for adolescents, with a significant percentage of adolescents either not receiving smoking prevention or cessation intervention or not recognizing or recalling receipt of intervention. The present study did not support fully previous research46 that found that pediatricians were more likely to encourage children and adolescents not to start smoking than to intervene with those who already smoke. Nonsmoking adolescents in the UC condition were not more likely to report that the pediatrician encouraged them in their efforts not to smoke than smoking adolescents were to report that the pediatrician intervened with them in their efforts to stop smoking. Indeed, providers in the UC condition were somewhat less likely to deliver the treatment steps recommended for nonsmokers, compared with smokers.
The finding that PEI scores were significantly greater for SI providers, compared with UC providers, supports strongly the effectiveness of the provider training and office systems in improving the delivery of smoking prevention and cessation intervention by pediatric providers to adolescents in the course of routine medical care. Pediatric providers in the SI condition spent only 1.4 minutes more, on average, discussing smoking during the visit, compared with providers in the UC condition, which suggests that this intervention may be feasible to implement in practice. The training was relatively brief and the office systems could be implemented easily in any pediatrician's office, with little additional demand on support staff members.
The majority of the pediatric providers in the present study were physicians, which precluded a meaningful analytic comparison of the SI and UC conditions according to type of provider. Although a review of the literature did not identify studies comparing smoking cessation interventions for adolescents across different provider types, studies comparing providers for other interventions found few, if any, differences between different types of providers. For example, in a survey of adult tobacco assessment and intervention practices and attitudes, primary care physicians, nurse practitioners/physician assistants, and public health nurses did not differ from one another and were more likely than chiropractors, dentists, and specialist physicians to report assessing, intervening, being supportive of cessation, and having the necessary tobacco skills and knowledge.70 In another study, the quality of HIV care provided by nurse practitioners and physician assistants was found to be similar to the care provided by physician HIV experts and better than the care provided by physician non-HIV experts.72 In a survey comparing nurse practitioners and physicians regarding weight management practices, nurse practitioners were more likely than physicians to obtain nutritional assessments and to discuss food labels, and physicians were more likely to recommend reduced-calorie diets and aerobic exercise, with no other significant differences being noted between the 2 provider types.73 Given the lack of significant differences found in other studies, it would be of interest to explore potential differences between pediatricians and other pediatric providers in the delivery of smoking prevention and cessation interventions to adolescents in future studies with larger samples of nonphysician providers.
In addition to SI condition, only a few adolescent characteristics were found to be associated with the delivery of intervention by pediatric providers. The finding that being a smoker was associated with a greater likelihood of reporting that the provider discussed cigarette smoking suggests that providers were more likely to address cessation with their adolescent smokers than they were to address prevention with nonsmokers. As mentioned earlier, this is in contrast to the findings from earlier studies, in which pediatricians reported being less likely to intervene with children and adolescents who already smoked and more likely to report encouraging those who did not smoke not to start.40 There are a number of possible reasons for the finding that adolescents who reported greater depression were less likely to report that their provider discussed smoking or, among nonsmokers/former smokers, were less likely to report that their provider congratulated them for not smoking and talked about nicotine addiction among friends who smoke. Adolescents experiencing depressive symptoms might have been less likely to recall what was discussed in the clinical encounter, thus underreporting the intervention received. Providers might have been less likely to conduct smoking interventions with adolescents who were depressed or exhibited depressive signs. The PEI items in which depression played a part among nonsmokers might be related to depressive symptoms. The first, congratulating the teen for not smoking, is a form of positive feedback, which depressed individuals are less likely to recognize and/or to accept. The second, having to do with discussing smoking among friends, might be attributable to depressed adolescents reporting that they have no friends who smoke, possibly because they have few or no friends, which results in the provider being less likely to discuss addiction among friends. The finding that providers were more likely to tell female adolescents that their discussion about smoking would be private suggests that providers may assume that female patients would be more concerned about confidentiality or that girls were more likely to recall discussion of confidentiality. Having a higher dependence score on the HONC was associated with a greater likelihood of the provider asking the adolescent smoker how much she or he smoked and asking whether she or he was interested in quitting, which suggests that these adolescents may present as more dependent, thereby prompting the provider to assess the amount smoked and to explore the motivation to quit.
A number of potential limitations to the present study should be noted. First, there is the question of the accuracy of adolescent recall of provider intervention behavior, particularly the recollection of specific topics discussed. Data from previous studies suggested that details regarding smoking interventions can be recalled accurately by adult patients74 and adolescents can recall whether or not smoking was discussed,66 but it is not known how well adolescents are able to recall the details associated with a discussion of smoking. Second, a systematic bias for reporting receipt of intervention among SI participants, compared with UC participants, might have occurred because of differences in the level of social desirability demand among the adolescents. Although randomization of sites to intervention versus UC should have controlled for this to some degree, social desirability was not assessed and could have been greater in the SI condition, because of the increased interaction of the adolescents with the intervention. Third, the small number of sites for randomization might have decreased the comparability of the 2 conditions. Last, a link between provider characteristics and PEI scores was not possible because of the need to ensure confidentiality of the provider surveys. There might be characteristics of providers beyond the intervention condition that would predict delivery of the intervention and PEI scores and that would be of interest to investigate.
A brief smoking prevention and cessation intervention for adolescents, based on existing clinical practice guidelines, can be delivered feasibly by pediatric providers in routine pediatric primary care clinical practice, with a high degree of fidelity. These findings demonstrate that pediatric providers in busy clinical practices can provide assistance to their adolescent patients regarding this critical health behavior with little additional time commitment. Given the tremendous potential for pediatricians and other pediatric clinicians to identify and to intervene with adolescents regarding tobacco, these findings are encouraging. As we continue to investigate how pediatric providers can best deliver efficacious smoking prevention and cessation interventions to their adolescent patients, it is recommended that measures such as the PEI be used to assess the feasibility of implementing these interventions.
We gratefully acknowledge support for this research from the National Cancer Institute (grant R01 CA80254-01A1).
We thank the research assistants who collected the data and the pediatric practices and their staff members who graciously took part in this study. To the adolescents who agreed to participate, we extend our deepest appreciation.
- Accepted March 31, 2006.
- Address correspondence to Lori Pbert, PhD, Division of Preventive and Behavioral Medicine, University of Massachusetts Medical School, 55 Lake Ave North, Worcester, MA 01655. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵US Department of Health and Human Services. Healthy People 2010: Understanding and Improving Health and Objectives for Improving Health. Washington, DC: US Government Printing Office;2000
- ↵DiFranza JR, Rigotti NA, McNeill AD, et al. Initial symptoms of nicotine dependence in adolescents. Tob Control.2000;9 :313– 319
- ↵US Department of Health and Human Services. Preventing Tobacco Use Among Young People: A Report of the Surgeon General. Atlanta, GA: Centers for Disease Control and Prevention, National Center of Chronic Disease Prevention and Health Promotion, Office on Smoking and Health;1994
- ↵Epps RP, Manley MW. A physician's guide to preventing tobacco use during childhood and adolescence. Pediatrics.1991;88 :140– 144
- ↵Centers for Disease Control and Prevention. Youth risk behavior surveillance: United States, 2003. MMWR Surveill Summ.2004;53(SS-2) :1– 96
- ↵National Institute on Drug Abuse. Monitoring the future: national results of adolescent drug use: overview of key findings, 2001. Available at: www.drugabuse.gov/PDF/overview2001.pdf. Accessed November 7, 2005
- ↵American Academy of Pediatrics, Committee on Substance Abuse. Tobacco's toll: implications for the pediatrician. Pediatrics.2001;107 :794– 798
- American Academy of Pediatrics, Committee on Substance Abuse. Tobacco, alcohol, and other drugs: the role of the pediatrician in prevention and management of substance abuse. Pediatrics.1998;101 :125– 128
- ↵American Academy of Pediatrics, Committee on Environmental Health. Environmental tobacco smoke: a hazard to children. Pediatrics.1997;99 :639– 642
- ↵American Academy of Pediatrics, Committee on Psychosocial Aspects of Child and Family Health. Guidelines for Health Supervision II. Elk Grove Village, IL: American Academy of Pediatrics;1988
- ↵Stein RJ, Haddock CK, O'Byrne KK, Hymowitz N, Schwab J. The pediatrician's role in reducing tobacco exposure in children. Pediatrics.2000;106 (5). Available at: www.pediatrics.org/cgi/content/full/106/5/e66
- ↵O'Connor EA, Hollis JE, Polen MR, Whitlock EP, Stevens VJ, Lichtenstein E. Expert system intervention for teen smoking in an HMO: an under-utilized opportunity. Presented at: annual meeting of the American Psychological Association; August 15–19, 1997; Chicago, IL
- Cohen SJ, Stookey GK, Katz BP, Drook CA, Christen AG. Helping smokers quit: a randomized controlled trial with private practice dentists. J Am Dent Assoc.1989;118 :41– 45
- ↵Fiore MC, Bailey WC, Cohen SJ, et al. Treating Tobacco Use and Dependence: Clinical Practice Guidelines. Rockville, MD: US Department of Health and Human Services;2000
- Elster A, Kuznets N. AMA Guidelines for Adolescent Preventive Services (GAPS): Recommendations and Rationale. Baltimore, MD: Williams & Wilkins;1994
- Green M. Bright Futures: Guidelines for Health Supervision of Infants, Children, and Adolescents. Arlington, VA: National Center for Education in Maternal and Child Health;1994
- US Preventive Services Task Force. Guide to Clinical Preventive Services. 2nd ed. Alexandria, VA: International Medical Publishing;1996
- US Department of Health and Human Services, Public Health Services, Office of Disease Prevention and Health Promotion. Put Prevention Into Practice. Washington, DC: US Government Printing Office; 1994
- American Academy of Family Physicians. AAFP Stop Smoking Program: Physician and Office Staff Manual. Kansas City, MO: American Academy of Family Physicians; 1987
- American Medical Association. Guidelines for Adolescent Preventive Services. Chicago, IL: American Medical Association; 1992
- Levenberg P, Elster A. Guidelines for Adolescent Preventive Services (GAPS): Recommendations and Rationale. Baltimore, MD: Williams & Wilkins; 1995
- ↵Epps RP, Manley MW. How to Help Your Patients Stop Smoking: A National Cancer Institute Manual for Physicians. Bethesda, MD: National Institutes of Health; 1991. NIH Publication 92–3064
- ↵Zapka JG, Fletcher K, Pbert L, Druker SK, Ockene JK, Chen L. The perceptions and practices of pediatricians: tobacco intervention. Pediatrics.1999;103 (5). Available at: www.pediatrics.org/cgi/content/full/103/5/e65
- Thorndike AN, Ferris TG, Stafford RS, Rigotti NA. Rates of US physicians counseling adolescents about smoking. J Natl Cancer Inst.1999;91 :1857– 1862
- ↵Centers for Disease Control and Prevention. Health care providers' advice on tobacco use to persons aged 10–22 years: United States, 1993. MMWR Morb Mortal Wkly Rep.1995;44 :827– 837
- ↵Kovacs M. Children's Depression Inventory. North Tonawanda, NY: Multi-Health Systems;1992
- ↵Hollis JF, Polen MR, Whitlock E, Velicer W, Redding CA, Licenstein E. Efficacy of a brief tobacco prevention and cessation program for teens seen in routine medical care (Teen Reach). Presented at: annual meeting of the Society of Behavioral Medicine; April 3–6, 2002; Washington, DC
- ↵Hollis JF, Polen MR, Whitlock EP, et al. Teen Reach: outcomes from a randomized, controlled trial of a tobacco reduction program for teens seen in primary medical care. Pediatrics.2005;115 :981– 989
- ↵Pratt CC, Katzev A, Henderson T, Ozetich MS. Building Results: From Wellness Goals to Positive Outcomes for Oregon's Children, Youth and Families : Salem, OR: Oregon Commission on Children and Families;1997
- ↵Centers for Disease Control and Prevention. Cigarette smoking. Available at: www.cdc.gov/nchs/datawh/nchsdefs/cigarettesmoking.htm. Accessed November 7, 2005
- ↵Prokhorov AV. Adolescent Nicotine Dependence Measure: Psychometrics, Reliability, and Biochemical Validation. Presented at: annual meeting of the Society of Prospective Medicine; October 25–28,1998; Newport, RI
- ↵DiFranza JR, Savageau JA, Rigotti NA, et al. Development of symptoms of tobacco dependence in youths: 30 month follow up data from the DANDY study. Tob Control.2002;11 :228– 235
- ↵Flint A, Young M, Druker S, Pbert L. The impact of a training program and experience with a novel provider-delivered adolescent smoking intervention on the confidence and practice behaviors of pediatric providers. Presented at: American Heart Association 44th Annual Conference on Cardiovascular Disease Epidemiology and Prevention; March 4–6,2004; San Francisco, CA
- ↵Pbert L. Provider and peer delivered youth smoking intervention: Air It Out. Presented at the Tobacco Control Investigators Meeting: Synthesizing Research for the Public's Health; June 2–4, 2004; San Diego, CA
- Copyright © 2006 by the American Academy of Pediatrics