Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004
To investigate family physicians' and pediatricians' practice of and perceived barriers to smoking cessation counseling among patients 18 years and younger.
Cross-sectional mail survey conducted between November 1, 1997, and January 31, 1998.
A stratified random sample selected from the 1997 American Medical Association Physician Masterfile of 1000 family physicians and pediatricians who practice in urban California, work at least 10% of the time in ambulatory care, and have at least 10% of patients 18 years and younger.
Main Outcome Measures
Physicians' adherence to 5 components of the National Cancer Institute's smoking cessation counseling recommendations (anticipate, ask, advise, assist, and arrange) and their perceived barriers to smoking cessation counseling.
A total of 429 physicians participated in the study. Physicians of both specialties were more likely to anticipate, ask, and advise patients about smoking than to assist with and arrange cessation activities. Family physicians were more likely than pediatricians to assist and arrange, including scheduling follow-up visits to discuss quitting (25.1% vs 11.7%; odds ratio [OR], 3.07; 95% confidence interval [CI], 1.22-7.73) and directing nursing staff to counsel patients (17.1% vs 10.9%; OR, 3.70; 95% CI, 1.30-10.60). The most common perceived barrier to counseling was the belief that children would provide inaccurate responses due to either the presence of parents (86.4%) or the fear that parents would be notified of their answers (74.0%). Pediatricians reported lack of counseling skills as a barrier to providing smoking interventions in greater proportion than did family physicians (24.9% vs 54.8%; OR, 0.29; 95% CI, 0.14-0.63; P<.001).
Improvement in smoking cessation counseling skills and practices is needed among physicians treating children and adolescents.
Most adult smokers in the United States began using tobacco during adolescence.1 It is well established that prolonged smoking can cause cancer and chronic respiratory and cardiovascular diseases in both men and women.2 Research also indicates that the earlier in life a child or adolescent tries smoking cigarettes, the more likely he or she is to become a regular smoker.3 The health consequences of cigarette smoking have been extensively documented for adults and youth. Therefore, preventing and reducing tobacco use among children and adolescents would provide long-term benefits for both the individual and society as a whole.
Routine visits to primary care physicians and pediatricians provide multiple opportunities for smoking cessation and prevention interventions for young smokers. Because physicians are viewed as credible health experts,4 young people may follow their advice more readily than that of other adults. Among adults, findings from numerous studies5- 7 have shown that brief physician tobacco cessation counseling increases cessation. Although its effectiveness in the prevention of tobacco use among youth has not yet been demonstrated, the National Cancer Institute (NCI) and several medical organizations8- 12 have recommended cessation training for physicians who care for children and adolescents. The NCI smoking cessation training program has 5 principal components: anticipate, ask, advise, assist, and arrange.8- 12 These are summarized as follows:
Anticipate: The physician addresses tobacco-related risks as they develop throughout the child's life.
Ask: The physician inquires about experimentation with smoking, starting when the patient reaches the age of 8 years.
Advise: The physician advises all children not to use tobacco products.
Assist: The physician provides the patient with self-help literature, referrals, or other mechanisms to help him or her quit smoking.
Arrange: The physician arranges for follow-up visits to reinforce quitting or to examine issues related to relapse.
Despite these recommendations, and although smoking among children and adolescents is an important public health issue, physicians often fail to assess and counsel patients regarding tobacco use. Limited information is available about physicians' ability to successfully identify smokers, counsel pediatric patients who smoke,13- 15 and assess the barriers to effective counseling. In addition, more information is needed to determine how a physician's specialty may affect the manner in which smoking prevention and cessation counseling is provided. We hypothesized that pediatricians are less aware than family physicians of cessation counseling techniques for dealing with smoking youth. In this study, we examined the self-reported delivery of the recommended smoking cessation components. In addition, we identified the perceived barriers to counseling children and adolescents among a random sample of family physicians and pediatricians in California.16 Specifically, we examined whether the physician's specialty influences the likelihood of delivering smoking cessation interventions and whether perceived barriers to counseling differ between specialties. Findings from this study will help to identify areas in which pediatricians may need additional training.
A stratified random sample of 1000 physicians (500 in each specialty) was selected from the 1997 American Medical Association (AMA) Physician Masterfile data set for California. Physicians were considered eligible if the AMA Physician Masterfile identified them as having a specialty in either family medicine or pediatrics in an urban area and being 65 years or younger. Based on responses provided to specific items in the questionnaire, the sample was further modified to include only those physicians who (1) were actively treating patients 18 years and younger, (2) practiced within ambulatory care settings at least 10% of the time, and (3) reported that at least 10% of their patients were children and/or adolescents.
A 31-question, self-administered questionnaire was developed based on current smoking cessation literature. The questionnaire was pretested with 10 practicing physicians to ensure readability, comprehension of the questions, and inclusion of all important subject matter. The survey was revised accordingly. In addition to background questions, the survey asked physicians about their practice of and perceived barriers to counseling children and adolescents on smoking-related issues. Survey packets mailed to physicians included a cover letter with contact telephone numbers, letters of support from professional organizations, the questionnaire, a stamped return envelope, and a $5 bill. Mailings began in November 1997, followed by a second mailing in January 1998. The University of California, San Francisco, Committee on Human Research approved all study protocols.
Background information was collected on sex, ethnicity (white, Asian/Pacific Islander, African American, Latino, or other group), age, country of birth (United States or other), smoking status (current, former, or never smoked regularly), year of graduation from medical school, country of medical training (United States or other), type of practice (private, solo, or group practice; community clinic; health maintenance organization; or academic practice), and percentage of time in ambulatory practice.
The survey asked physicians about their understanding of the smoking initiation process. Physicians were asked, "In your opinion, what is the average age for children to begin (1) experimenting with smoking tobacco (trying to smoke at least one full cigarette), and (2) smoking tobacco regularly (smoking at least once a month)?" Response categories were ages 6 through 8, 9 through 11, 12 through 14, 15 through 17, and older than 17 years.
The survey contained questions pertaining to the 5 components of NCI's smoking cessation counseling recommendations: anticipate, ask, advise, assist, and arrange.
Anticipate and Ask. Physicians were asked to estimate what percentage of their patients (in age groups ≤8, 9-10, 11-12, 13-14, 15-16, and 17-18 years) they ask about the experimental or regular use of tobacco. The response categories were on a continuous spectrum from 0% to 100% and anchored at multiples of 10. For purposes of analysis, we subsequently categorized the responses as above or below a certain cutoff point.
Advise. Physicians were asked to estimate the average number of minutes (0, 1-2, 3-5, 6-10, or ≥11 minutes) per visit they spend providing young smokers with smoking cessation counseling. In addition, physicians were asked to consider their young patients (≤18 years) who experiment with tobacco or smoke tobacco regularly and indicate the percentage of these patients with whom they conduct smoking cessation activities, such as (1) advise patients to quit, (2) record in medical chart as a problem, (3) counsel on the health risks of smoking, (4) discuss smoking effects on personal appearance, (5) address possible weight change issues, (6) assess motivation and interest in quitting, and (7) ask if patients are willing to set a quit date. As described herein, the response categories were on a continuous spectrum from 0% to 100% and anchored at multiples of 10.
Assist. Physicians were asked to estimate the percentage of their child or adolescent patients who experiment with tobacco or smoke tobacco regularly that they (1) have nursing staff counsel to stop smoking and (2) provide with pamphlets on smoking cessation. The response categories for these items were also on a continuous spectrum from 0% to 100% and anchored at multiples of 10.
Arrange. Physicians were asked about the percentage of their child or adolescent patients who experiment with tobacco or smoke tobacco regularly for whom they (1) schedule follow-up visits to discuss quitting and (2) telephone to check on progress in cessation. The response categories for these items were on the same continuous spectrum from 0% to 100% and anchored at multiples of 10.
Responses to these questions on physicians' practice of smoking cessation activities were dichotomized into 2 groups: physicians who reported conducting the cessation activity with at least 70% of their patients and those who reported conducting the cessation activity with less than 70% of their patients. This cutoff point was selected based on the goal of 75% cited in the US Department of Health and Human Services' Healthy People 2000 objectives.17 Due to the design of our data collection instrument, we were limited to choosing either 70% or 80% as a cutoff point; we selected 70% because it was less restrictive.
Barriers to Counseling Youth Who Smoke. Physicians were given a list of 12 potential barriers to the delivery of smoking cessation services and asked to state the extent to which they agreed that these barriers affected their cessation counseling practices. Barriers included were time constraints, physician attitudes toward the effectiveness of cessation services, reimbursement issues, lack of smoking cessation counseling skills, perceived anger of parents, inaccuracy of self-report, and inappropriateness of addressing smoking issues with minors. Responses to questions on perceived barriers to the delivery of smoking cessation services were collapsed into 2 categories: agree some or a lot and agree a little or not at all.
Descriptive statistics were computed separately and combined for the physician specialties (family medicine and pediatrics) (SAS/Stat User's Guide, version 8; SAS Institute Inc, Cary, NC). The main dependent variables were the smoking cessation activities, dichotomized using a 70% cutoff point as described herein.
Bivariate analyses were performed to explore differences between the 2 medical specialties using the χ2 test for categorical data and the t test (2-tailed, unpaired) for continuous data. Based on differences found in the bivariate analyses, multivariate logistic regression models were constructed to assess differences between family physicians and pediatricians with respect to their practice of and perceived barriers to providing smoking cessation services. Odds ratios (ORs), 95% confidence intervals (CIs), and significance levels were calculated. Analyses were controlled for by sex, ethnicity (white vs other), age (continuous), country of birth (United States vs other), smoking status (ever smoked regularly vs never smoked regularly), year of graduation from medical school (≤1980 vs >1980), country of medical training (United States vs other), and percentage of work time spent in ambulatory care settings (continuous).
From the original sample of 1000 physicians, 101 of the returned surveys were deemed ineligible because the physician's current practice did not match study criteria (eg, retired, not practicing in ambulatory care, or not seeing children or adolescent patients), the physician's specialty differed from the AMA Physician Masterfile listing, or survey packets were deemed undeliverable by the post office. Of the remaining 899 physicians presumed eligible, 499 returned completed questionnaires for a response rate of 55.5%. Respondents did not differ significantly from nonrespondents in regard to sex (P = .32), specialty (P = .26), or ethnicity (P = .53). The sample was further reduced to include only those physicians whose practice consisted of at least 10% children or adolescents. This resulted in a final sample of 429 physicians for this report.
Table 1 shows that, relative to family physicians, a higher percentage of pediatricians were women (44.2% vs 28.9%, P<.001). No significant differences between physician specialties were found for ethnicity, age distribution, country of birth, smoking status, year of graduation, country of medical training, type of practice, or time spent in ambulatory care. Most physicians (92.0%) believed that adolescents try cigarettes for the first time by the age of 14 years, and 55% thought adolescents initiate regular smoking by the age of 14 years (data not shown). Pediatricians indicated a younger age of smoking initiation than family physicians (Figure 1). For example, 53.2% of the pediatricians indicated that the age of smoking experimentation was between 12 and 14 years of age compared with 48.1% of the family physicians (P = .01). The same relationship was seen in the assessment of the initiation of regular smoking. More than 50% of pediatricians (55.2%) indicated that adolescents begin smoking regularly between the ages of 12 and 14 years compared with 45.4% (P = .02) of family physicians.
Physicians' assessment of patients' age at start of smoking experimentally and smoking regularly (California, 1997-1998).
This component was measured as the percentage of patients in various age groups who physicians reported asking about the experimental or regular use of tobacco products (Figure 2). Overall, both groups of physicians reported asking a greater proportion of patients in the older age groups than in the younger. Family physicians reported asking a significantly higher proportion of patients 10 years and younger about their smoking habits than did pediatricians (11.4% vs 5.8% for those 8 years and younger, P = .01; 19.3% vs 13.6% for those aged 9-10 years, P = .03). However, a significantly higher proportion of pediatricians reported asking those 13 to 14 years old about their tobacco use than did family physicians (67.8% vs 59.8%, respectively, P = .03).
Physicians' report of percentage of young patients (≤18 years) asked about experimental or regular tobacco use by age group and physician specialty (California, 1997-1998).
The survey assessed both the techniques physicians used in counseling smoking patients and the time they spent counseling. The amount of time reported by physicians in counseling their young patients about smoking cessation is presented in Figure 3. Overall, there were no significant differences between the 2 specialties regarding counseling time. Most physicians, regardless of specialty type (90.0% of pediatricians, 83.2% of family physicians), reported providing patients with 1 to 5 minutes of smoking cessation counseling. Also, half of all physicians (48.6% of family physicians, 52.6% of pediatricians) reported spending longer counseling sessions (3 minutes or more) with their young patients who smoke.
Percentage of physicians reporting time spent counseling young patients (≤18 years) who smoke by physician specialty (California, 1997-1998).
Table 2 presents information on several behaviors recommended for promoting smoking cessation among adolescents. These behaviors are related to the advise, assist, and arrange components of the physician interventions and are ranked by reported prevalence. More than three fourths of the physicians reported implementing each of the following counseling practices with at least 70% of smoking patients: advising patient to quit, 92.1%; counseling on health risks, 89.1%; and recording smoking behavior within patients' medical charts, 76.7%.
More than half (65.6%) of the physicians reported that they assess patients' motivation and interest in quitting with at least 70% of their patients, and nearly half (49.2%) discuss the effects that smoking has on personal appearance. Addressing possible weight change issues as they relate to smoking was reported by more than one third (36.1%) of the physicians surveyed, and just less than one third (31.5%) reported distributing smoking cessation pamphlets to their patients. More than one fourth (28.0%) of the physicians reported that they encouraged smokers to set a quit date. The least acknowledged behaviors were scheduling follow-up visits (17.6%) and directing nursing staff to counsel patients (13.7%). Less than 5% of the physicians reported contacting at least 70% of their young patients who smoke to check on their cessation progress.
The 2 specialty groups counseled their patients in different proportions. Family physicians reported addressing possible weight change issues (OR, 3.67; 95% CI, 1.66-8.16; P<.001), setting a quit date (OR, 2.41; 95% CI, 1.08-5.40; P = .03), scheduling a follow-up appointment to discuss quitting (OR, 3.07; 95% CI, 1.22-7.73; P = .02), and directing the nursing staff to counsel the patient about quitting (OR, 3.70; 95% CI, 1.30-10.60; P = .02) more often than did pediatricians.
Table 3 presents information regarding physicians' perceived barriers to smoking cessation counseling. More than 50% of all the physicians cited the following barriers to successful smoking cessation counseling: the assessment would be inaccurate with parents present, 86.4%; the patient will inaccurately self-report out of fear that his or her parents will be notified, 74.0%; counseling is time-consuming, 61.0%; and counseling would anger parents who are present at the visit, 52.3%. As many as 38.3% of physicians indicated that cost was a barrier to smoking cessation practices with their pediatric patients, and 30.7% believed that these programs would be ineffective in younger patients. More than 40% of the physicians indicated that they do not have the skills required to counsel children and adolescents on smoking cessation. A significantly higher proportion of pediatricians (54.8%) than family physicians (24.9%) indicated that lack of smoking cessation counseling skills was a barrier (P<.001). A small proportion of physicians reported that smoking cessation counseling was not appropriate for younger patients because children are legally too young to smoke (13.6%) or that inquiring about their smoking behavior would seem to condone it (12.0%).
With respect to asking patients about their smoking behavior, the lowest proportion of reported counseling (5.8%) was with the youngest patients (≤8 years of age) and the highest (81.9%) with the oldest age group (17-18 years). Even acknowledging some degree of overreporting due to self-report bias, these rates fall short of the recommendation to ask all adolescents about smoking.11,18 Other studies15,19 previously conducted in the Northeast have found higher rates. For example, Franzgrote et al15 indicated that physicians reported screening younger adolescents during more than 71% of routine visits, whereas Klein et al19 reported that 91% of physicians ask adolescents about smoking. Nevertheless, the proportions observed in our study are somewhat higher than those found in other research based on adolescents. For example, 1993 data from the Morbidity and Mortality Weekly Report20 show that only 25% of adolescents and young adults reported that a clinician had ever discussed cigarette smoking with them. The 2002 study of adolescents by Alfano et al21 reported that only 43% of adolescents were ever asked by their physician whether they smoke.
Further examination of the anticipating and asking behavior by age group indicates that physicians are not asking about smoking during the age at which children and adolescents are most likely to try their first cigarette, which typically ranges from 11 to 12 years.22 This finding is consistent with other studies.15,23,24 One possible explanation is that physicians may not view children and adolescents as susceptible to trying cigarettes. However, physicians' self-report in this study does not support this explanation, since most physicians believed that adolescents try cigarettes for the first time before the age of 14 years and that many adolescents are regular smokers by the age of 17 years. Despite this understanding, the physicians only achieved the desired counseling rate of 70% or more (as recommended by the US Department of Health and Human Services) for those patients 15 years and older. Although pediatricians believe that initiating cigarette smoking occurs at a significantly younger age than do family physicians, this belief is not demonstrated in their counseling behavior. In fact, pediatricians' proportion of young patients asked about tobacco use is significantly lower than that of family physicians for the 2 youngest age groups. However, in the 13- to 14-year-old age group, pediatricians' proportion is significantly higher than that of family physicians.
Among the specific counseling behaviors, physicians of both specialties tended to report a higher prevalence of behaviors in the advise component of the counseling intervention, such as "advise patient to quit," "counsel on health risks of smoking," and "record in medical chart as a health problem." In contrast, the behaviors less cited were those classified as the arrange and assist components of the intervention, such as "assist with self-help literature," "give pamphlets on smoking cessation," "ask if willing to set a quit date," "schedule follow-up visit to discuss quitting," and "direct nursing staff to counsel patient to stop smoking." These results are consistent with other studies25,26 of smoking cessation counseling of adults in which physicians tended to ask and advise but were less likely to fulfill the assist or arrange components of the intervention.
Family physicians were more likely to report performing the advise and arrange components of the intervention than were pediatricians. One plausible explanation is that family physicians tend to have a greater familiarity with smoking cessation counseling procedures than physicians in other specialties,27 because they use such procedures with their adult patients and can apply them to their younger patients. In contrast, pediatricians do not have this same experience or counseling training, as indicated by the pediatricians in our sample. They cited a lack of counseling skills as a barrier in greater proportion than family physicians. This reported lack of confidence in their counseling skills is troublesome, given the importance of preventing smoking uptake at earlier ages, and may be due in part to a deficiency in formal training. A recent study28 shows that, although medical schools provide students with epidemiological information about smoking, they do not incorporate smoking cessation guidelines into their curricula and do not provide adequate training for clinical interventions. Based on our findings, we would recommend that medical training programs incorporate smoking cessation training for all specialties dealing with adults and children and that smoking cessation training programs be added to the curricula of medical schools in both the preclinical and postclinical years.
There were no differences between the 2 physician groups with regard to the least cited counseling procedure—telephoning patients to check their progress with quitting—which falls into the arrange category. This procedure may be difficult to implement in busy pediatrics or family medicine practices. In our sample, "time-consuming" was the third most commonly cited barrier to counseling among all physicians.
The most cited barriers related to issues of confidentiality. Physicians seemed concerned that patients might provide inaccurate responses to inquiries about their smoking habits because their parents are present or because they do not trust the physician to keep their responses confidential. These issues of confidentiality are found throughout the adolescent literature.29- 33 In our study, the lower proportion of pediatricians asking about smoking in the younger age groups may be related to these concerns. Because parents are more likely to accompany younger children into the examination room, physicians may be less inclined to discuss smoking-related issues with these patients.
Both physicians and their young patients view issues of confidentiality as being a large barrier to honest communication. Hence, physicians must make a concerted effort to diminish their patients' fears. Adolescents have reported a greater willingness to communicate with those physicians who assure confidentiality.29 However, many physicians do not broach the subject of confidentiality with their young patients.29 The simple act of assuring confidentiality, coupled with an effort to inquire about smoking-related behaviors in privacy, may increase the likelihood of conducting successful cessation interventions. The literature supports this concern; adolescents indicate that worries about privacy in clinical settings may decrease their eagerness to seek health care for sensitive health problems and may hinder their communication with clinicians.
In research on adult populations, a lack of organizational support and lack of financial incentives for counseling smokers have also been cited as barriers to smoking cessation counseling.26,34 In our study, nearly 40% of physicians echoed this concern by indicating that inadequate reimbursement poses a barrier to delivery of smoking cessation practices to their young patients.
These data confirm results of prior studies and expand on them by examining the independent components of the NCI physician-based counseling recommendations. Overall, we found that physician assessment of smoking should be initiated with patients of a younger age. Physicians seem to be successful in advising their patients to quit smoking; however, there seems to be less of an effort to include the complementary tasks of counseling, such as assisting patients by providing further information and arranging follow-up visits. Issues of privacy and confidentiality seem to play a role in deterring counseling and should be addressed in the implementation of smoking cessation interventions.
One limitation of this study is that results are based on physicians' self-reported behaviors, which were not validated by patients, their guardians, or medical chart reviews. The reported rates are therefore likely to overestimate actual counseling behavior by physicians. Furthermore, the strict tobacco control policies in California compared with other states and regions present the possibility that these findings may have limited generalizability. Finally, this study only investigated the smoking-related counseling that children and adolescents received from pediatric and family physicians, because the selection process limited the scope of the study to primary care physicians in California. Future research should include other types of clinicians (eg, nurses, nurse practitioners, or physician assistants) who play a large role in children's and adolescents' health care.
In summary, physicians are not currently providing their young patients with desired levels of smoking cessation counseling. Physicians' belief that their patients might be concerned about confidentiality and inaccurately report smoking behaviors and physicians' feeling that they lack appropriate counseling skills appear to be large deterrents to counseling. Physician education and training in regard to physician-patient confidentiality and smoking cessation counseling are needed at all levels. In addition, the focus of physician-based interventions should be not only on the epidemiological aspects of smoking behavior but also on improving smoking counseling training for physicians of all primary care specialties and for pediatricians in particular.
Although most adult smokers in the United States begin smoking during adolescence, physicians often fail to provide smoking cessation counseling to their pediatric and adolescent patients. This study aims to describe smoking cessation counseling practices and perceived barriers to counseling among physicians of 2 specialties that interface with youth smokers: family physicians and pediatricians. In particular, it addresses the paucity of research on physicians' performance of the NCI's recommended smoking cessation counseling components (anticipate, ask, advise, assist, and arrange) among young patients. This study found that physicians are not currently providing pediatric and adolescent patients with desired levels of smoking cessation counseling. Consistent with previous studies of adults, the data showed that physicians are more likely to ask and advise young patients about smoking than to assist with or arrange cessation and follow-up activities. To address the main perceived barriers to counseling and increase its practice, physicians should be encouraged to assure their patients of the confidentiality of their responses and should receive formal training in counseling skills.
Corresponding author: Eliseo J. Pérez-Stable, MD, Department of Medicine, Division of General Internal Medicine, University of California, San Francisco, 400 Parnassus Ave, Room A-405, Box 0320, San Francisco, CA 94143 (e-mail: email@example.com).
Accepted for publication July 14, 2003.
This research was supported by grant 6RT-0368 from the University of California Tobacco-Related Disease Research Program.
We thank Sylvia Correro, JD, and Sarah Pelta, MA, for their assistance.
Kaplan CP, Pérez-Stable EJ, Fuentes-Afflick E, Gildengorin V, Millstein S, Juarez-Reyes M. Smoking Cessation Counseling With Young PatientsThe Practices of Family Physicians and Pediatricians. Arch Pediatr Adolesc Med. 2004;158(1):83-90. doi:10.1001/archpedi.158.1.83