|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Brief Physician-Initiated Quit-Smoking Strategies for Clinical Oncology Settings: A Trial Coordinated by the Eastern Cooperative Oncology GroupFrom the Fox Chase Cancer Center, Philadelphia, PA; Dana-Farber Cancer Institute, Boston, MA; Duluth Clinic, Duluth, MN; Park Nicollet Health Services, St. Louis Park, MN; and AMC Research Center, Denver, CO. Address reprint requests to Robert A. Schnoll, PhD, Psychosocial and Behavioral Medicine Program, Fox Chase Cancer Center, 510 Township Line Rd, Cheltenham, PA 19012; email: RA_Schnoll{at}fccc.edu.
Purpose: Although tobacco use by cancer patients increases the risk of relapse, diminishes treatment efficacy, and worsens quality of life, about one third of patients who smoked before their diagnosis continue to smoke. Because patients have regular contact with oncologists, the efficacy of a physician-based smoking cessation treatment was evaluated. Methods:Cancer patients (n = 432) were randomly assigned to either usual care or a National Institutes of Health (NIH) physician-based smoking intervention. The primary outcome was 7-day point prevalence abstinence at 6 and 12 months after study entry. Results: At the 6-month follow-up, there was no significant difference in quit rates between the usual care (11.9%) and intervention (14.4%) groups, and there was no significant difference between the usual care (13.6%) and intervention (13.3%) groups at the 12-month follow-up. Patients were more likely to have quit smoking at 6 months if they had head and neck or lung cancer, began smoking after the age of 16, reported at baseline using a cessation self-help guide or treatment in the last 6 months, and showed greater baseline desire to quit. Patients were more likely to have quit smoking at 12 months if they smoked 15 or fewer cigarettes per day, had head and neck or lung cancer, tried a group cessation program, and showed greater baseline desire to quit. Finally, there was greater adherence among physicians to the NIH model for physician smoking treatment for patients in the intervention versus the usual care group. Conclusion: While training physicians to provide smoking cessation treatment to cancer patients can enhance physician adherence to clinical practice guidelines, physician smoking cessation interventions fail to yield significant gains in long-term quit rates among cancer patients.
THE RANGE of adverse health consequences of tobacco use are well documented and widely acknowledged. However, there is generally less awareness of the important ways that continued tobacco use by cancer patients can further compromise the patients’ health. Studies with cancer patients show that continued smoking after diagnosis and treatment reduces survival time,1–3 increases the risk of a recurrence or a second primary tumor,4–6 reduces treatment efficacy,7 and exacerbates and prolongs treatment-induced complications such as mucositis, dry mouth, loss of taste and voice, impaired pulmonary function and wound healing, and tissue and bone necrosis.8–12 Despite the significant increase in risk for these negative health effects, upwards of one quarter to one third of cancer patients who smoked before their diagnosis continue to do so after diagnosis.13–19 Such findings lend strong support for the need to integrate formal smoking cessation clinical services within Comprehensive Cancer Centers. Over the last 20 years, there has been a steady accumulation of research on the efficacy of smoking cessation interventions provided in a medical context.20,21 One treatment approach that has received a sizable amount of research attention is physician-based intervention.22 This approach, whereby the physician is trained to deliver cessation advice and assistance, yields abstinence rates of 10% to 11%.22–24 The effectiveness of this intervention approach may be attributable to the fact that hospitalizations or visits to physicians represent a "window of opportunity" during which patients are particularly motivated to quit smoking and are especially receptive to smoking cessation treatments.20,21,25 Indeed, patients highly respect physicians’ opinions and recommendations, and they enter a medical encounter with a focus on their health and a readiness to engage in actions relevant to improving their well-being, such as quitting smoking.26 The growth of physician-based smoking interventions has been spurred by a number of additional factors as well, including the finding that more than 90% of smokers who are able to achieve abstinence have not used formal clinical programs to do so27 and the observation that this approach to treating nicotine addiction can reach a substantial number of smokers, because more than 80% of smokers visit a physician at least once each year.28 When considering the recent emphasis on assessing the effectiveness of smoking treatments in terms of their impact value (ie, the abstinence rate yielded by the intervention multiplied by the participation rate),29,30 physician-based smoking interventions represent a relatively potent way to address the epidemic of tobacco use in the United States. In light of accumulated data supporting the effectiveness of physician-based smoking interventions, public health agencies have urged physicians to implement smoking interventions by providing treatment guidelines and training manuals.31 One model for physician-based smoking treatments, referred to as the "5 A’s," was devised by the Agency for Healthcare Research and Quality:32 Ask (screen for tobacco use), Advise (provide a personalized and strong quit message), Assess (evaluate the smoker’s readiness to quit), Assist (provide cessation counseling, pharmacotherapy, and self-help guides), and Arrange (appraise cessation progress). Because use of this model can yield a 10% to 11% quit rate and upwards of 80% of smokers could be reached by this treatment approach, physician-based smoking treatments have the potential to yield sizable reductions in the overall population rate of tobacco use. To date, however, the potential effectiveness of this model for reducing continued tobacco use by cancer patients has yet to be explored. A review of the literature33 uncovered only five studies of smoking treatments with cancer patients.34–38 Further, the studies that have been conducted have used small samples and/or have failed to maintain sufficient internal validity.33 As such, despite the need to provide cancer patients with formal cessation assistance,39 cancer patients who smoke remain a group of smokers for whom smoking interventions have yet to be adequately evaluated. Thus, in this study, we examined the effectiveness of a 5 A’s, physician-based smoking cessation intervention for promoting long-term abstinence among cancer patients. We expected that patients who smoke and who are randomly chosen to receive a physician-based smoking intervention would show higher quit rates at 6-month and 12-month follow-up time points, versus patients who smoke who receive usual care. In addition, we explored whether sociodemographic (eg, age, sex), medical (eg, cancer site or stage), and behavioral (ie, desire to quit, expectations about quitting) factors, along with smoking history (eg, level of nicotine addiction, number of years smoked), predict long-term abstinence. With few exceptions,13,40 there has not been sufficient assessment of factors predictive of long-term tobacco use among cancer patients, although such analyses can provide important information for the design of smoking cessation interventions with this population.33 Finally, we explored the degree to which the physician-based intervention enhanced physician adherence to the 5 A’s model (eg, improved the rate at which physicians assessed smoking behavior) to provide a possible explanation for an effect of the physician-based intervention on patient quit rates.
Participants This study was originally conducted by Fox Chase Cancer Center (Philadelphia, PA) outside a cooperative. From December 1990 to May 1991, institutions that were members of the Eastern Cooperative Oncology Group (ECOG), the Southwestern Oncology Group, and the Cancer and Leukemia Group B participated in the study (see Table 1  ), although not through cooperatives. ECOG assumed direction of the study from June 1992 until it was closed in June 1999. Participants were 19 years of age or older; were diagnosed with stage I–II cancer (of any type) or stage III–IV breast, prostate, or testicular cancer or lymphoma; and had an ECOG performance status score of 0 or 1. To be eligible, patients must have given informed consent, have been scheduled to return to the treatment site, and have reported smoking one or more cigarettes in the past 30 days or have self-identified as a smoker. Four hundred thirty-five patients were randomly assigned; three patients were dropped (two because they were ineligible and one because of a duplicate registration), leaving the final sample at 432.
Procedure Participating physicians at the recruitment sites identified potentially eligible patients, using standard intake evaluation forms. Eligible patients had the study objectives explained to them, and they were presented with an informed consent form. Consenting patients then completed a baseline assessment, which included an evaluation of their smoking history and certain behavioral factors that are strongly related to the ability to quit smoking (eg, desire to quit). Sociodemographic and medical data were recorded from medical charts or self-report. Random assignment to the usual care condition or the physician-based intervention was conducted using permuted blocks with dynamic balancing within the recruitment sites. Patients randomly assigned to the usual care arm (n = 218) received standard care after recruitment. Because physicians, from an ethical standpoint, could not withhold quit advice and prevent the patient from seeking cessation treatment outside the study, patients in the usual care arm might have been provided with physician quit advice and/or may have sought assistance with quitting. At the end of the study, patients in this arm who were smoking were instructed to call 1–800-4-CANCER for assistance and/or referral to a cessation program. Patients randomly assigned to the physician-based intervention (n = 217) received quit advice and assistance during a physician visit in accordance with the National Institutes of Health guidelines for physician-based smoking interventions;31 all patients, regardless of the site of their tumor, received the same intervention. For the intervention, the physician provided brief (< 5 minutes) quit advice, which emphasized the overall benefits of quitting (eg, improved general health) and the specific health benefits for cancer patients (eg, lower chance for recurrence); tried to convince the patient to set a quit date and worked with the patient to identify the quit date; discussed nicotine replacement therapy (NRT) with the patient and provided a prescription if indicated (ie, for patients who exhibited high nicotine dependence); provided the patient with a self-help smoking cessation guide (ie, Freedom from Smoking for You and Your Family)41 and the 1–800-4-CANCER telephone number for additional assistance and/or referral to a cessation program; and assessed cessation progress during subsequent medical visits. ECOG-trained data coordinators assisted physicians with standardizing intervention components at each site. All patients were contacted by telephone to complete a 6-month and a 12-month assessment, which contained the same information as the baseline assessment and questions (at the 6-month assessment only) to assess physician adherence to the smoking cessation treatment model (eg, whether or not the physician discussed the patient’s smoking habits).
Measures Medical data. Data on the patient’s diagnosis and health were collected from the medical chart or from the patient. These included cancer site, time since diagnosis (ie, the duration from diagnosis to baseline assessment), cancer stage, ECOG performance status, and physical symptoms (eg, trouble breathing, pain or tightness in chest) and diseases or conditions (ie, diabetes, high blood pressure, heart trouble/attack, emphysema, chronic bronchitis, or asthma). Smoking history. Participants indicated age of smoking initiation; average number of cigarettes smoked per day in the past 7 days; how soon they smoke after they wake (to assess level of nicotine addiction); whether others in the household smoke; number of lifetime 24-hour quit attempts; whether or not they had tried to quit in the past year; whether they had been advised to quit by a health care worker; whether they had ever tried individual counseling, group counseling, or NRT to quit smoking; whether, in the past 6 months, they had tried to reduce the amount they smoked, tried to quit, or used a quit-smoking guide; whether they used other forms of tobacco besides cigarettes (eg, cigars); and whether they are seriously contemplating quitting smoking in the next 6 months. Behavioral data. Two attitudinal variables were assessed as potential predictors of quitting, because in previous studies these have been linked to the ability to quit smoking.40 First, using six Likert-type questions, which ranged from 0 (not at all) to 4 (very much), desire to quit smoking was assessed (eg, "How much, if at all, do you want to quit smoking?"). Second, using six Likert-type questions, ranging from 0 (not at all) to 4 (very much), negative expectations about quitting were assessed (eg, "How much of a problem do you think being irritable, angry, or tense might be for you when you quit smoking?"). For each of these, a sum score was formed and used as a predictor variable in the analyses; higher scores on each composite variable represent, respectively, greater desire to quit or greater negative expectations about quitting. Smoking status. This variable, which served as the primary outcome measure, was assessed using 7-day point prevalence abstinence, as has been done in previous studies.14 The patient was considered abstinent if he or she reported not smoking for at least 1 week before the interview (ie, 7-day point prevalence abstinence). This measure of smoking status is as reliable and valid as other common measures of smoking status, such as 24-hour point prevalence, 30-day prolonged abstinence, and 6-month prolonged abstinence.42 All nondeceased nonrespondents were considered to be smokers at the follow-up assessments. Physician adherence to the 5 A’s model. At the 6-month follow-up, patients reported physician adherence to the physician-based smoking cessation clinical guidelines (ie, the 5 A’s). A yes or no format was used to assess whether physicians discussed smoking with the patient, suggested that the patient quit smoking, reviewed with the patient the benefits of quitting, helped the patient to set a quit date, sent a follow-up letter about the patient’s quitting plans, provided the patient with quit-smoking materials, and provided a prescription for NRT.
Statistical Analyses
Second, a Bayesian model was used to assess baseline predictors of long-term abstinence at the follow-up, including all participants regardless of missing data in the covariates. A Bayesian approach was used because this would allow for including all the subjects in the model, even those that have missing data in the covariates. There was a fairly high level of missing data (see Table 2
Finally, to explore the degree to which the intervention affected physician adherence to the 5 A’s model, we calculated the proportions across the study arms of patients who indicated that the physician assisted them with quitting smoking. Because this aim was primarily exploratory in nature, we restrict these analyses to descriptive analysis.
Descriptive sociodemographic, medical, and smoking history data for the sample are shown in Tables 2  and 3. (Behavioral data are not shown in a table because there were no differences across conditions.) There were only two significant differences in patient characteristics across the study arms (P < .05): there were more males in the usual care arm than in the physician-based intervention (39% v 29%, respectively); and a lower proportion of patients in the usual care arm reported experiencing heart trouble or a heart attack in the past year than did patients in the physician-based intervention (3.7% v 11.6%, respectively). Because neither of these variables was related to long-term quit rates, they were not included as covariates in subsequent analyses.
The Effect of Physician-Based Smoking Intervention on Long-Term Quit Rates
The quit rates at 12 months are shown in the bottom of Table 4 Finally, there were no significant differences in quit rates at 6 months and at 12 months when the results were stratified by sex. At 6 months, the quit rates for males were 12.5% in the usual care arm and 15.0% in the physician-based intervention arm (P = .34); at 12 months, the quit rates for males were 13.6% in the usual care arm and 14.8% in the physician-based intervention arm (P = .46). For females, the 6-month quit rates were 11.0% in the usual care arm and 13.1% in the physician-based intervention arm (P = .45); the 12-month quit rates were 13.6% in the usual care arm and 9.8% in the physician-based intervention arm (P = .34).
Examination of Predictors of Long-Term Smoking Cessation
Table 5
The Bayesian multiple regression model of 12-month quitting is shown in Table 5
Physician Adherence to Smoking Cessation Treatment Guidelines
When taken together, our analyses yield several important findings. First, contrary to our hypothesis, we did not find that the provision of a physician-based smoking intervention increased long-term quit rates among cancer patients. This finding is consistent with a study by Gritz et al,36 which reported no significant difference in the rates of quitting between patients provided with a physician-based cessation treatment and patients randomly assigned to a control comparison group. However, our finding contradicts the general literature concerning the effect of physician-based smoking interventions, which indicates that brief physician-based interventions yield upwards of a 10% to 11% increase in quit rates (versus usual care).22
There are several possible explanations for our null finding. First, our finding may be the result of methodological confounding that was caused by the inability to use a "true" control group for this study (ie, an arm that received no active treatment). Although our findings indicated greater adherence to the 5 A’s model among the physicians in the intervention condition (Table 6 A second possible explanation for our null finding is that the brief, relatively low-intensity, physician-based intervention evaluated here was not sufficient to increase long-term quit rates among cancer patients, who may be considered "hard-core" smokers because they continue to smoke despite their diagnosis. The relatively low overall rate of long-term quitting across both conditions (13.2% and 13.4% at the 6-month and 12-month follow-ups, respectively), when compared with previous intervention studies with cancer patients, lends support for this explanation. For example, in a study that evaluated the benefits of a nurse-managed, 5 A’s model–based smoking intervention for lung cancer patients, Browning et al34 found that 10 of 14 (71%) patients in the intervention condition were abstinent at a 6-month follow-up, versus six of 11 (55%) patients in the usual care group. Compared with the quit advice and self-help literature provided in the usual care condition, the intervention condition in the study by Browning et al34 consisted of eight structured clinical sessions involving quit advice, guidance in the establishment of a quit date, teaching of behavioral and cognitive strategies to manage stress, provision of support and encouragement, use of self-help guides, recommendation to use NRT and bupropion, and follow-up telephone booster contacts. That the regression models in the present study found that patients who used more-intensive forms of smoking treatment (ie, group counseling, NRT) were more likely to have quit smoking in the long term supports this explanation as well. Thus, future studies in this area may need to assess higher-intensity interventions (ie, physician-based intervention plus adjunctive behavioral and pharmacologic treatment), with the understanding that the brief, low-impact, physician-based intervention that may be sufficient for initiating quitting among healthy smokers (even in a medical context) is not adequate for smokers diagnosed with cancer. A third possible explanation for a lack of statistical impact of the intervention relates to the inclusion of patients whose primary cancer was not linked to tobacco use (eg, breast or testicular cancer, lymphoma), because the intervention, in large part, relies on the high level of quit motivation among the individuals diagnosed with a tobacco-related illness. The inclusion of such patients, who may have lower levels of desire to quit because they do not necessarily perceive their current health problem as linked to their tobacco use, may have contributed to the low cessation rates across the two conditions and created a scenario from the beginning of the study that was to make a treatment effect unlikely. Supporting this explanation is the higher cessation rate demonstrated in prior smoking intervention studies that used either head and neck or lung cancer patients.34,36,37 In addition, the present analyses found that cancer site was a strong predictor of long-term quit rates, with patients who have head and neck or lung tumors (whose etiology is strongly linked to tobacco use) exhibiting significantly higher quit rates than patients diagnosed with other forms of cancer. Thus, it may be the case that physician-based interventions such as the one tested here are effective only with patients manifesting a tobacco-related illness. A final possible explanation for the null findings concerns the outcome measure selected. In this study, we did not assess immediate quit rates after cessation advice but, rather, focused on long-term abstinence. It may be that the physician intervention yielded a significant increase in initial quitting after the provision of cessation advice but was not effective at promoting long-term maintenance of abstinence. (However, we did not assess immediate quit rates after quit advice.) This distinction between initial quitting and long-term abstinence is important to consider, because if a physician-based treatment can motivate initial cessation but fail to promote long-term abstinence, low-cost relapse prevention techniques, rather than incorporating more intensive interventions at the outset, may be a useful adjunct to a physician’s quit advice. A second important finding from this study concerns the results from the prediction models of long-term quitting. These models identified several medical, smoking, and behavioral factors (assessed at baseline) that are linked to the patient’s ability to quit smoking in the long term. As mentioned, the tumor site has an effect on the likelihood of quitting; patients diagnosed with head and neck or lung cancer were substantially more likely to be abstinent at the 6-month and 12-month follow-up, versus patients with tumors at other sites. This finding may be attributable to the awareness among patients that the etiology of head and neck and lung tumors is strongly linked to tobacco use. This recognition, in turn, may spark greater quit motivation. Indeed, a study with head and neck cancer patients found that attributing the cause of the disease to tobacco use predicted the ability to quit smoking after diagnosis.44 Alternatively, smoking rates may be lower among head and neck cancer patients because of the greater likelihood of physical disability from treatments for this cancer type, which makes it impossible to smoke. Regardless, to boost quit motivation and eventual cessation, more-intensive interventions may need to be targeted toward patients with tumors that are not traditionally linked with tobacco use. In addition, interventions may need to be targeted to patients with a longer time since diagnosis, because these patients were less likely to have been abstinent at the 12-month follow-up, as reported previously.38,40 After the diagnosis, risk of smoking may increase over time because, initially, debilitating medical treatments or symptoms make it impossible for patients to smoke. Alternatively, once sufficient time has elapsed, and treatments have been completed, patients may minimize the seriousness of their diagnosis, and they might be, thus, less committed to maintaining abstinence from tobacco use. Finally, smoking rates may be lower at the outset as a result of a strong message from a physician requiring that the patient abstain from smoking during treatment, but as time elapses, the effect of this message fades. Regardless of the explanation, this finding highlights the need for relapse prevention for patients. Four variables, assessed at baseline, related to the patient’s smoking history predicted long-term quitting in the regression models. Patients who started smoking before the age of 16 (ie, smoked for a greater number of years) and patients who smoked 15 cigarettes per day or more (ie, heavier smokers) were less likely to quit at follow-up. The link between smoking duration and the amount smoked, on the one hand, and the ability to quit smoking, on the other hand, has been shown in large prospective studies with noncancer populations.45,46 In addition, patients who reported using a quit-smoking guide or treatment in the past 6 months or who tried a group cessation program were more likely to have quit smoking in the long term. This finding is also supported by literature that has demonstrated the benefits of such treatments.32,47 These results replicate those previously found concerning smoking history and the ability to quit, and they indicate that patients who start smoking at a younger age or patients that are heavy smokers require more intensive cessation treatments and that methods for promoting use of cessation interventions should be explored to increase the likelihood that patients will abstain from tobacco use in the long term. Finally, patients who showed a strong desire to quit smoking at baseline were more likely to be abstinent in the long term. Prospective studies show that the desire to quit smoking (often referred to as quit motivation, readiness to change, or stage of change) is a consistent predictor of smoking cessation.36,48 Further, treatment studies show that initial motivation to quit smoking predicts success in a smoking intervention.49 Quit motivation predicted a better outcome in treatment studies with cancer patients, as well.36,38 Such data sparked the development of behavior change theories, such as the Transtheoretical model,50 and clinical approaches, such as motivational interviewing.51 One important contribution of this research has been the matching of clinical techniques to the individual’s level of quit motivation.52 For instance, although action-oriented techniques (eg, NRT, behavioral counseling) can promote quitting among those motivated to quit, such methods may be met with resistance by those with low quit motivation. Instead, unmotivated smokers require a unique set of techniques designed to boost motivation and prepare the person for engaging in action-oriented strategies.52 Thus, as seen in the general population of smokers, interventions for cancer patients who smoke may also need to be tailored to the person’s level of readiness to quit to maximize the impact of treatment. A third important finding concerns the effect of the intervention on physician adherence to the 5 A’s model. Despite the potential benefits of physician-based smoking interventions, only 20% to 25% of U.S. smokers receive smoking cessation counseling from their physician.53 One explanation for the low rate of physician adherence to the 5 A’s model is the lack of training in smoking cessation counseling for physicians. Ferry et al54 surveyed 120 U.S. undergraduate medical programs and found that only 2.4% of those schools require students to complete a course in tobacco education. Because our study found that physicians in the intervention condition showed greater adherence to the 5 A’s model, providing physicians with basic training in conducting smoking interventions may improve physician practice. One caveat to highlight is sources of bias in patient recall of physician practices, which may yield unreliable measures of physician behavior. First, because these measures were taken 6 months after physician intervention, patients may have forgotten that physicians had provided such treatment, thereby artificially lowering the rate of adherence. Second, patients’ recall may have been influenced by whether they had made a quit attempt and failed; these patients may be reluctant to report physician attempts to encourage their cessation because they want to provide a rationale for their inability to maintain abstinence. Our findings should be viewed in the context of additional limitations. First, because the use of self-reporting, versus biochemical verification, to determine smoking status can be unreliable,55 a certain number of patients may have been incorrectly classified at the follow-ups. It is important to note that large studies have shown false reporting rates to be about 3% to 5%,56,57 which indicates that, if deception occurred in this study, the rate was small. Nevertheless, future studies that evaluate smoking interventions with cancer patients should use biochemical verification of smoking status. Second, the external validity of this study is limited by the sample’s lack of ethnic diversity. Even though patient ethnicity has not been found to be associated with the patient’s ability to quit smoking after diagnosis, future studies in this area should use samples that better approximate the ethnic distribution of U.S. cancer patients. Third, as mentioned, a "true" control group was not used. Thus, confounding the content of from exposure to smoking cessation treatment among patients in the usual care arm may have undermined the study’s internal validity. Fourth, practical constraints required that patients be randomly assigned, rather than physicians. Thus, the same physicians implemented the usual care as well as the physician-based intervention. If this methodological approach affected the data, it would be expected to yield a type I error (ie, declaring the physician-based intervention to be effective when it is not), because physicians’ knowledge of which treatment condition the patients are in could influence their treatment approach in favor of the study hypothesis. With a null finding for treatment condition, this cannot be the case (indeed, only a type II error is plausible). Thus, although we acknowledge the methodological shortcoming of having the same physician provide both treatments, we do not believe that this would explain the null findings. Finally, the eligibility criteria used allowed recent quitters to be included in the study. This procedure can artificially increase quit rates. However, as a product of the random assignment procedures, the proportion of these individuals in each condition could be reasonably expected to be equivalent. With these limitations acknowledged, this study makes an important contribution to the literature on smoking treatments for cancer patients, because only one previous study36 tested a physician-based smoking cessation with cancer patients. As the recognition for the need to address tobacco use among cancer patients grows, studies such as this can help guide the implementation of clinical initiatives. It is important that this study suggests the need for more intensive smoking interventions for this population than basic physician-based treatments, as well as the need for the development of relapse prevention strategies. Although the provision of physician advice and basic assistance (eg, self-help guides, referral for NRT) should be seen as the standard of care (and likely to be effective in a primary care setting), a substantial proportion of cancer patients (especially those with non-tobacco-related neoplasms and those in latter stages of recovery) are likely to need more intensive cessation treatment, which could include the use of antidepressants and formal behavioral counseling tailored to the patient’s quit motivation,33 or long-term relapse prevention in the form of quit lines or support groups. The seamless integration of more intensive smoking interventions or relapse prevention approaches within the oncologic context represents an important priority for Comprehensive Cancer Centers and potentially a promising approach to assisting patients to achieve a better quality of life.
This ECOG-coordinated study (E1Y92; Robert L. Comis, MD) was supported by the National Cancer Institute, National Institutes of Health, Department of Health and Human Services grants CA23318, CA66636, CA21115, CA27525, and CA13650. Its contents are the author’s responsibility and do not necessarily represent the views of the National Cancer Institute.
1. Goodman MT, Kolonel LN, Wilkens L, et al: Smoking history and survival among lung cancer patients. Cancer Causes Control 1:155–163, 1990[CrossRef][Medline] 2. Silverman S, Gorsky M, Greenspan D: Tobacco usage in patients with head and neck carcinomas: A follow-up study on habit changes and second primary oral/oropharyngeal cancers. J Am Dent Assoc 106:33–35, 1983[Abstract] 3. Stevens MH, Gardner JW, Parkin JL, et al: Head and neck cancer survival and lifestyle change. Arch Otolaryng 109:746–749, 1983
4. Day GL, Blot W, Shore R, et al: Second cancers following oral and pharyngeal cancers: Role of tobacco and alcohol. J Natl Cancer Inst 86:131–137, 1994 5. Kawahara M, Ushijima S, Kamimori T, et al: Second primary tumours in more than 2-year disease-free survivors of small-cell lung cancer in Japan: The role of smoking cessation. Br J Cancer 78:409–412, 1998[Medline]
6. Tucker MA, Murray N, Shaw EG, et al: Second primary cancers related to smoking and treatment of small-cell lung cancer. Lung Cancer Working Cadre. J Natl Cancer Inst 89:1782–1788, 1997
7. Browman GP, Wong G, Hodson I, et al: Influence of cigarette smoking on the efficacy of radiation therapy in head and neck cancer. N Engl J Med 328:159–163, 1993
8. Dresler CM, Roper C, Patterson GA, Cooper JD: Effect of physician advice on smoking cessaton in patients undergoing thoracotomy. Chest 110:1199–1202, 1993 9. Des Rochers C, Dische S, Saunders MI: The problem of cigarette smoking in radiotherapy for cancer in the head and neck. Clin Oncol (R Coll Radiol) 4:214–216, 1992[CrossRef] 10. Karim AB, Snow GB, Siek HT, et al: The quality of voice in patients irradiated for laryngeal carcinoma. Cancer 51:47–49, 1983[CrossRef][Medline] 11. Rugg T, Saunders ML, Dische S: Smoking and mucosal reactions to radiotherapy. Br J Radiol 63:554–556, 1990[Abstract] 12. Whittet HB, Lund JV, Brockbank M, et al: Serum cotinine as an objective marker for smoking habit in head and neck malignancy. J Laryngol Otol 105:1039, 1991 13. Gritz ER, Schacherer C, Koehly L, et al: Smoking withdrawal and relapse in head and neck cancer patients. Head Neck 21:420–427, 1999[CrossRef][Medline] 14. Ostroff JS, Spiro RH, Kraus DH: Prevalence and predictors of continued tobacco use after treatment of patients with head and neck cancer. Cancer 75:569–576, 1995[CrossRef][Medline] 15. Spitz MR, Fueger JJ, Chamberlain RM, et al: Cigarette smoking patterns in patients after treatment of upper aerodigestive tract cancers. J Cancer Educ 5:109–113, 1990[Medline] 16. Vander Ark W, DiNardo LJ, Oliver DS: Factors affecting smoking cessation in patients with head and neck cancer. Laryngoscope 107:888–892, 1997[CrossRef][Medline]
17. Fujisawa T, Iizasa T, Saitoh Y, et al: Smoking before surgery predicts poor long-term survival in patients with stage I non-small cell lung cancer. J Clin Oncol 17:2086–2091, 1999 18. Gritz ER, Nisenbaum R, Elashoff RE, et al: Smoking behavior following diagnosis in patients with stage I non-small cell lung cancer. Cancer Causes Control 2:105–112, 1991[CrossRef][Medline]
19. Sridhar KS, Raub WA: Present and past smoking history and other predisposing factors in 100 lung cancer patients. Chest 101:19–25, 1992 20. Halpern MT, Schmier JK, Ward KD, et al: Smoking cessation in hospitalized patients. Respir Care 45:330–336, 2000[Medline] 21. Orleans CT, Kristeller JL, Gritz ER: Helping hospitalized smokers quit: New directions for treatment and research. J Consult Clin Psychol 61:778–789, 1993[CrossRef][Medline] 22. Ockene JK, Zapka JG: Physician-based smoking intervention: A rededication to a five-step strategy to smoking research. Addict Behav 22:835–848, 1997[CrossRef][Medline] 23. Glynn TJ: Physicians and a smoke-free society. Arch Intern Med 148:1013–1016, 1988[CrossRef][Medline] 24. Kottke T, Battista RN, DeFriese GH, et al: Attributes of successful smoking cessation interventions in medical practice. A meta-analysis of 39 controlled trials. JAMA 259:2883–2889, 1988[Abstract] 25. Emmons KM, Goldstein MG: Smokers who are hospitalized: A window of opportunity for cessation interventions. Prev Med 21:262–269, 1992[CrossRef][Medline]
26. Stein RJ, Haddock CK, O’Byrne KK, et al: The pediatrician’s role in reducing tobacco exposure in children. Pediatrics 106:1–17, 2000 27. US Department of Health and Human Services: The Health Benefits of Smoking Cessation: A Report of the Surgeon General (DHHS Publication No. 90-8416). Washington, DC, U.S. Government Printing Office, 1990 28. US Department of Health and Human Services: Current Estimates From the National Health Interview Survey, 1993 (Series 10: Data from the National Health Survey No. 190) (DHHS Publication No. 95-1518). Hyattsville, MD, Public Health Service, Centers for Disease Control, National Center for Health Statistics, 1994 29. Abrams DB, Orleans CT, Niaura RS, et al: Integrating individual and public health perspectives for treatment of tobacco dependence under managed care: A combined stepped-care and matching model. Ann Behav Med 18:290–304, 1996[Medline] 30. Velicer WF, Prochaska JO: An expert system intervention for smoking cessation. Patient Educ Couns 36:119–129, 1999[CrossRef][Medline] 31. Glynn TJ, Manley MW: How to Help Your Patients Stop Smoking (National Institutes of Health Publication No. 98-3064). Bethesda, MD, Public Health Service, National Institutes of Health, 1998 32. Fiore M, Bailey W, Cohen S, et al: Treating Tobacco Use and Dependence. Clinical Practice Guideline. Rockville, MD, US Department of Health and Human Services, Public Health Service, 2000 33. Schnoll RA, Lerman C: Smoking behavior and smoking cessation among head and neck cancer patients, in Ensley JF, Gutkind JS, Jacobs JR, et al (eds): Head and Neck Cancers. San Diego, CA, Academic Press (in press) 34. Browning KK, Ahijevych KL, Ross P, et al: Implementing the Agency for Healthcare Policy and Research’s smoking cessation guideline in a lung cancer surgery clinic. Oncol Nurs Forum 27:1248–1254, 2000[Medline] 35. Griebel B, Wewers ME, Baker CA: The effectiveness of a nurse-managed minimal smoking-cessation intervention among hospitalized patients with cancer. Oncol Nurs Forum 25:897–902, 1998[Medline] 36. Gritz ER, Carr CR, Rapkin D, et al: Predictors of long-term smoking cessation in head and neck cancer patients. Cancer Epidemiol Bio Prev 2:261–270, 1993[Abstract] 37. Stanislaw AE, Wewers ME: A smoking cessation intervention with hospitalized surgical cancer patients: A pilot study. Cancer Nurs 17:81–86, 1994[Medline]
38. Cox LS, Patten CA, Ebbert JO, et al: Tobacco use among patients with cancer treated for nicotine dependence. J Clin Oncol 20:3461–3469, 2002 39. Gritz ER: Smoking and smoking cessation in cancer patients. Br J Addict 86:549–554, 1991[CrossRef][Medline] 40. Schnoll RA, Malstrom M, James C, et al: Correlates of tobacco use among smokers and recent quitters diagnosed with cancer. Patient Educ Couns 46:137–145, 2002[CrossRef][Medline] 41. Strecher VJ, Rimer BK, Monaio KD: Development of a new self-help guide—Freedom from smoking for you and your family. Health Educ Q 16:101–112, 1989[Medline] 42. Velicer WF, Prochaska JO: A comparison of 4 smoking cessation outcome measures. Presented at Soc Behav Med Annual Meeting, Seattle, WA, March 21–24, 2001 43. Spiegelhalter D, Thomas A, Best N, et al: Bayesian inference using Gibbs sampling. MRC Biostatistics Unit, Institute of Public Health, Cambridge, UK. http://www.mrc-bsu.cam.ac.uk/bugs/ 44. Christenson AJ, Moran PJ, Ehlers SL, at al: Smoking and drinking behavior in patients with head and neck cancer: Effects of behavioral self-blame and perceived control. J Behav Med 22:407–418, 1999[CrossRef][Medline] 45. Murray RP, Gerald LB, Lindgren PG, et al: Characteristics of participants who stop smoking and sustain abstinence for 1 and 5 years in the Lung Health Study. Prev Med 30:292–400, 2000 46. Nides MA, Rakos RF, Gonzales D, et al: Predictors of initial smoking cessation and relapse through the first 2 years of the Lung Health Study. J Consult Clin Psychol 63:60–69, 1995[CrossRef][Medline] 47. Engstrom PE, Clapper M, Schnoll RA, et al: Prevention of tobacco-related cancers, in Bast RC, Kufe D, Pollock RE, et al: (eds): Cancer Medicine (5th edition). Hamilton, ON, Canada, BC Decker, 2000, pp 127–140
48. Osler M, Prescott E: Psychosocial, behavioural, and health determinants of successful smoking cessation: A longitudinal study of Danish smokers. Tobacco Control 7:262–267, 1998 49. Sciamanna C, Hoch J, Duke G, et al: Comparison of five measures of motivation to quit smoking among a sample of hospitalized smokers. J Gen Intern Med 15:16–23, 2000[CrossRef][Medline] 50. Prochaska JO, DiClemente CC, Norcross JC: In search of how people change: Applications to addictive behaviors. Am Psychol 47:1102–1114, 1992[CrossRef][Medline] 51. Miller WR, Rollnick S: Motivational Interviewing: Preparing People to Change Addictive Behaviors. New York, NY, Guilford Press, 1991 52. Rollnick S, Mason P, Butler C: Health Behavior Change. Edinburgh, UK, Churchill Livingstone, 1999
53. Thorndike AN, Rigotti NA, Stafford RS, et al: National patterns in the treatment of smokers by physicians. JAMA 279:604–608, 1998
54. Ferry LH, Grissino LM, Runfola PS: Tobacco dependence curricula in US undergraduate medical education. JAMA 282:825–829, 1999 55. Boyd NR, Windsor RA, Perkins LL, et al: Quality of measurement of smoking status by self-report and saliva cotinine among pregnant women. Matern Child Health J 2:77–83, 1998[CrossRef][Medline]
56. Wells AJ, English PB, Posner SF, et al: Misclassification rates for current smokers misclassified as nonsmokers. Am J Public Health 88:1503–1509, 1998 57. Glasgow RE, Mullooly JP, Vogt TM, et al: Biochemical validation of smoking status: Pros, cons, and data from four low-intensity intervention trials. Addict Behav 18:511–527, 1993[CrossRef][Medline] Submitted April 18, 2002; accepted September 19, 2002. This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
N (0,106). To allow proper imputation of missing values, prior distributions for the covariates were assumed to be Bernoulli (P), with P distributed as Uniform (0,1). The BUGS program (Bayesian inference using the Gibbs sampling algorithm) was used to estimate the coefficients and 95% credible intervals.
.25 in the univariate analysis, one at a time. Variables were eliminated if their 95% credible intervals included the null value and their odds ratios (ORs) were lower than 1.5; variables with ORs above 1.5, even if their 95% credible intervals include the null, likely have clinical significance. To check the assumption that the sampling and nonresponse mechanisms are jointly ignorable, we compared the distribution of the response variable according to the presence or absence of missing data in the covariates, and we found no indication of nonignorable missing data; P values are two-sided unless stated otherwise. 
