Offering telephone counseling to smokers using pharmacotherapy

Nicotine & Tobacco Research Volume 7, Supplement 1 (April 2005) S19–S27

Offering telephone counseling to smokers using pharmacotherapy Raymond G. Boyle, Leif I. Solberg, Stephen E. Asche, Jackie L. Boucher, Nicolaas P. Pronk, Catharine J. Jensen [Received 10 June 2004; accepted 3 December 2004]

Whereas telephone-based counseling has been found to be effective in supporting smokers interested in quitting smoking, it is not known whether proactive efforts to reach smokers receiving cessation medications will enhance their likelihood of successful quitting. We had an opportunity to test, in a health plan setting, an offer of telephonebased counseling with smokers identified from health plan records as recently filling a prescription for nicotine replacement therapy or bupropion. After we removed 31 members determined to be ineligible, 1,329 were randomly allocated to receive an invitation either to telephone-based counseling (n5663) or to a control group (n5666). On average, 7 days (range53–15 days) elapsed from the day of the prescription fill until the Center for Health Promotion began calling to invite members to participate in telephone counseling. The Center for Health Promotion was able to reach 49% of those in the intervention group (323/663). Of these members, 118 (37%) declined any participation. Therefore, in response to the proactive contact, 63% (205/323) of those reached and 31% (205/663) of those eligible participated in some smoking cessation counseling. At the 3-month follow-up, we observed an increased quit rate (33.1% vs. 27.4%) among health plan members randomized to telephone-based smoking cessation counseling. The results varied by gender and amount smoked. In addition, the variables associated with quitting in a multivariate logistic regression model included older age and using more than 30 days of medication.

Introduction The evidence-based U.S. Public Health Service clinical practice guideline for tobacco use highlighted several significant advances in the support available to smokers who want to quit smoking (Fiore et al., 2000). These advances included the introduction of new pharmacotherapies for tobacco cessation and the delivery of counseling for tobacco dependence through telephone contact. Telephone-based quitlines for smoking cessation can be organized to provide proactive calls, but most of these quitlines rely on spontaneous or solicited calls from smokers (Lichtenstein, Glasgow, Lando, Ossip-Klein, & Boles, 1996). In a proactive Raymond G. Boyle, Ph.D., Leif I. Solberg, M.D., Stephen E. Asche, M.A., HealthPartners Research Foundation, Minneapolis, MN; Jackie L. Boucher, M.S., Nicolaas P. Pronk, Ph.D., Catharine J. Jensen, B.A., Center for Health Promotion, HealthPartners, Minneapolis, MN. Correspondence: Raymond Boyle, Ph.D., M.P.H., HealthPartners Research Foundation, PO Box 1524, MS#21111R, Minneapolis, MN 55440-1524 USA. Tel: +1 (952) 967-5227; Fax: +1 (952) 967-5022; E-mail: [email protected]

approach, a counselor initiates calls to smokers to provide support for their quit attempts. Several variations of this proactive model are in use in more than 30 states and multiple countries around the world (Ossip-Klein & McIntosh, 2003). A possible extension of this approach is to provide telephone support as an adjunct for smokers who have obtained evidence-based cessation medications. A few studies have evaluated this approach in a variety of populations, including low-income smokers (Solomon, Scharoun, Flynn, Secker-Walker, & Sepinwall, 2000; Wadland, Soffelmayr, & Ives, 2001), smokers recruited through primary care, hospitals, or health plans (Lando et al., 1997; Ockene et al., 1991; Simon, Carmody, Hudes, Snyder, & Murray, 2003), and smokers recruited from the general population (Macleod, Charles, Arnaldi, & Adams, 2003; Reid, Pipe, & Dafoe, 1999). The present study extends this line of research by examining the effect of providing telephone counseling as an adjunct to filling prescriptions for nicotine replacement therapy (NRT) or bupropion for health

ISSN 1462-2203 print/ISSN 1469-994X online # 2005 Society for Research on Nicotine and Tobacco DOI: 10.1080/14622200500078048

S20

OFFERING TELEPHONE COUNSELLING TO SMOKERS

plan members. The studied health plan has two existing but independent cessation support programs: a pharmacy benefit for cessation medications that does not require cessation counseling, and a telephone-based smoking cessation counseling program that has relied mostly on spontaneous calls from smokers. In this paper we report the short-term outcomes of a randomized controlled trial of a strategy to offer telephone-based cessation counseling to health plan members who smoked and who did or did not have smoking-related chronic diseases. We included a study arm of smoking members with chronic diseases because other studies have shown that these patients are significantly more likely to be counseled by physicians and because cessation is particularly important for them (Jaen, Crabtree, Zyzanski, Goodwin, & Stange, 1997; Jaen, Stange, Tumiel, & Nutting, 1998; Thorndike, Rigotti, Stafford, & Singer, 1998; Ossip-Klein et al., 2000). We hypothesized that, compared with a control group, providing proactive telephone counseling would increase use of the medication, quit attempts, and smoking cessation rates among health plan members filling a prescription for covered smoking cessation medications. In addition, we hypothesized that proactive telephone counseling would benefit smokers with a chronic disease compared with members without a chronic disease.

Method Setting The present study was conducted at HealthPartners, a member-governed, not-for-profit, managed care organization located in Minneapolis, Minnesota. HealthPartners provides health insurance to approximately 700,000 members, two-thirds of whom receive care in contracted clinic systems. HealthPartners has had a smoking cessation pharmacy benefit since 2001. Under this benefit, 6 months of smoking cessation medication is available to members. The covered medications include bupropion, nicotine patches, nicotine gum, and nicotine inhaler. Members are required to obtain a prescription from their health care provider and to make a nominal copay to use the benefit but are not reimbursed for or required to receive counseling. The health plan contracts with a national pharmacy benefit manager (PBM) to service pharmacy claims for members. At the time of the study initiation, the health plan was already contracting with the PBM to mail a followup resource letter to members filling a prescription for bupropion. When a prescription was filled for a member, the PBM sent the member a one-page letter on HealthPartners letterhead. The purpose of the letter was to (a) congratulate the member on his or

her efforts to quit smoking and (b) describe additional health plan and community resources available to assist the member’s smoking cessation efforts. The present study was designed to evaluate a new health promotion program conducted by the HealthPartners Center for Health Promotion (CHP). CHP had an existing protocol for offering proactive telephone counseling to members who completed health risk appraisals or who were referred by their health care providers. However, most counseling recipients were those who called in spontaneously after reading about the program in health plan member materials. In an effort to increase access to cessation counseling, CHP decided to expand their proactive calling protocol to include members who filled a prescription for a smoking cessation medication. To evaluate this expanded protocol, members filling prescriptions were randomly selected to be offered telephone counseling. Three months later, investigators from HealthPartners Research Foundation surveyed equal numbers of members offered counseling and not offered counseling. This project was reviewed and approved by the HealthPartners Institutional Review Board. Resource letter The first step of this project was to modify the pharmacy resource letter to inform members that the health plan might contact them directly to discuss health plan services for smoking cessation. In addition, the PBM was asked to expand the list of members receiving the letter by including prescriptions for nicotine replacement products (gum, patch, and inhaler) in the member letter protocol. However, at the time the resource letter was expanded to include NRT products, several generic patch products were mistakenly omitted by the PBM. As a result, a smaller proportion of members with nicotine patch prescription fills was included in the study than otherwise would have been the case. Sample and procedures The PBM provided a weekly file of all health plan members who had been mailed a resource letter in the previous week. A total of 3,099 members were mailed a resource letter by the PBM between March 2002 and April 2003. Using this list of members, we searched the health plan administrative data for claims-based ICD9 codes for smoking-related chronic diseases. Members could have more than one chronic disease. Among those with any chronic disease, we included the following: diabetes (17%), hypertension (47%), coronary heart disease (10%), hyperlipidemia (40%), chronic respiratory diseases

NICOTINE & TOBACCO RESEARCH

(39%), and other arterial/vascular diseases (9%). An equal number of members without any of these chronic diseases was included. Figure 1 presents the study profile. A total of 1,360 smokers were randomly allocated to receive an invitation to either telephone-based counseling (n5679) or a control group (n5681). However, during the follow-up survey we identified 31 members who were ineligible for the study. The reasons for exclusion included use of tobacco other than cigarettes (n520; cigars, snuff, or chewing tobacco), denial of receiving a prescription for smoking (n54), use of bupropion for depression (n53), and member death (n54). The final study sample was, therefore, 1,329. Ineligible members were distributed equally between the two groups. Intervention For the present study, CHP dedicated a health educator trained in smoking cessation to make telephone calls as a follow-up to the pharmacy resource letter. The purpose of the telephone call was to invite members to participate in either general counseling (i.e., unstructured counseling focused on tobacco-cessation-related topics of member interest)

Figure 1.

S21

or a telephone-based course for smoking cessation (i.e., structured course covering specific topics known to be effective for tobacco cessation). When individuals enrolled in the smoking cessation course, they received a course manual and could participate in up to nine follow-up telephone counseling sessions. During the course, participants identified personal smoking behaviors, completed activities to prepare for a quit date, selected a quit date, and practiced skills to manage high-risk situations after quitting. Members who chose the general counseling option worked through cognitive behavior change activities similar to those in the telephone-based course; however, they did it without the structure of the smoking cessation course. General counseling is often preferred by members because it is less structured and more informal but is still tailored to the individual. There was no charge for either option. Each week, half of the sample of patients with and without a chronic disease identified as filling a prescription was randomly selected to be called and invited to participate in counseling. The protocol was the same as other CHP proactive calls and involved three attempts to reach a member on different days and times, followed by an ‘‘unable to reach’’ letter that asked the member to call into the call center.

Study design. NRT5nicotine replacement therapy; PBM5pharmacy benefit manager.

S22


Accompanying the letter was an educational brochure to reinforce the member’s tobacco cessation efforts. All calls were followed through an existing referral and tracking system. We identified only two members (0.6%) who called in response to the ‘‘unable to reach’’ letter. This rate is close to CHP’s previous experience. The specific goals of the invitation call were to

N N N N

Assess member questions or concerns related to the prescription for bupropion or NRT Assess smoking status and interest in discussing plans for smoking cessation Describe the counseling and tobacco cessation course available through the call center Engage interested members in either general counseling or the telephone-based course for smoking cessation

Based on the outcome of the invitation call, we classified members as unable to reach, reached but declined to participate, participated in general counseling, or enrolled in the tobacco cessation course. Follow-up survey The follow-up survey was conducted by the Data Collection Center within the HealthPartners Research Foundation, using staff not involved in the intervention. Health plan members in both the comparison group and the invitation treatment group were contacted 3 months after the fill date of their prescription and asked to complete a follow-up survey (n51,360). The protocol for data collection followed a modified Dillman method (Dillman, 2000). The first survey mailing was followed by a postcard reminder to all subjects after 1 week, and a second survey was sent to nonrespondents 3 weeks later (Dillman, 2000). If participants did not return a completed survey after the two mailings, we made up to 10 attempts to reach the participant by phone (at various days and times) to administer the survey. The four-page, 27-item questionnaire was composed largely of items tested previously in other studies. It requested information about the smoker and smoking behaviors as well as questions about the clinician encounter when the pharmacotherapy prescription was obtained. Included were questions about satisfaction with the health plan, quitting attempts, use of various quit aids, the duration of medication used, and the number of days of use. Age and gender were obtained from health plan records. Tobacco cessation was determined based on two questions: The first question was ‘‘Have you smoked a cigarette, even a puff in the last 7 days?’’ The second question asked if they had used any other tobacco products in the last 7 days. No biochemical validation of self-reported abstinence was performed.

Statistical power This evaluation was powered based on a withinstrata comparison (in which one strata has a chronic disease and the other does not) to be able to detect a 10% difference in self-reported quit rates (15% vs. 25%) between the invitation to telephone counseling group and the control group. To detect this difference with 80% power using a Type 1 error rate of 0.05 and a two-sided test of proportions, we needed 250 people per group (proactive call vs. control). Assuming a loss to follow-up of 25%, the recruitment target was 667 people in each of the two chronic disease strata. Analysis Because the follow-up period was relatively short, we considered 7-day point-prevalent quitting a primary outcome. Days of medication use among those who used the medication (,30 vs. 30+) and quit attempts (0 vs. 1+) were secondary outcomes. We report here separate bivariate analyses testing each dependent variable with a list of covariates including age, gender, education, marital status, chronic disease status (present, absent), amount smoked (less than a pack per day, more than a pack per day), and years smoked. Contingency tables, chi-square tests, and logistic regression were used to assess the associations. We conducted a multivariate logistic regression analysis to predict quit status from the intervention, while controlling for covariates. The covariates used in this analysis included items having at least marginally significant (p,.10) bivariate associations with quitting status (age, duration of medication used, amount smoked) and items needed for the testing of key interaction terms (gender, chronic disease status, amount smoked at the time of prescription fill). To test whether the relationship between intervention group and quit status varied by these items, three interaction terms (intervention6 gender, intervention6chronic disease status, intervention6amount smoked at baseline) were entered in a second block of the analysis as separate equations. The final model retained the intervention variable, covariates, and significant interaction terms. Analyses were conducted using SPSS v. 10.

Results Proactive contact Of the 663 eligible members randomized to the intervention group, we obtained complete CHP contact data on 658 patients. On average, 7 days (range53–15 days) elapsed from the day of the prescription fill until CHP began calling to invite members to participate in telephone counseling.


Using the proactive telephone call protocol, CHP was able to reach 49% (323/663) of members in the intervention group. Of these members, 118 (36%) declined any participation. Therefore, in response to the proactive contact, 63% (205/323) of those reached and 31% (205/663) of those eligible participated in some smoking cessation counseling. The majority were interested only in general counseling (n5147), whereas another 58 joined the telephone-based cessation course (see Figure 1). At the 3-month follow-up we contacted all subjects randomized into the study. Completed questionnaires were obtained from 1,035 of 1,360 study participants (completion rate578%, adjusted for ineligibles). Of these participants, 533 had a chronic disease (51%). We found no statistically significant differences in response to the survey by intervention condition, type of medication filled, gender, or chronic disease status. However, older members were more likely to respond than were younger members (linear trend, p,.003). The demographic and tobacco use variables of responders are reported in Table 1. Respondents were, on average, 47 years of age (range519–82 years). For the purposes of analysis, we categorized respondents as either younger than age 50 years, or 50 years or older. Most respondents were female, married, and educated beyond high school. They self-reported smoking an average of 21 cigarettes per day (SD512) at about the time they filled the prescription, and they had smoked for an average of 25 years (SD513). In addition, members with a chronic disease were more likely than those without a chronic disease to be older (mean age551.6 vs. 41.7, p,.001), have more years smoking (M529.4 vs. 21.1, p,.001), smoke more cigarettes per day (M522.2 vs. 20.3, p5.012), and use NRT (28.7% vs. 18.7%, p,.001), and they were less likely to be college educated (64.4% vs. 71.6%, p5.013). Use of the quit aid Overall, 86.1% of those surveyed reported actually using the cessation medication they had obtained, with no difference by intervention status, type of medication, gender, or chronic disease status. On average, 53% used the medication within 2 days of filling the prescription and 96% used it within 30 days. No difference was found in when members started using medication after filling the prescription by type of medication or chronic disease state. Respondents also were asked how long they had used the medication. Smokers randomized to the intervention group were significantly more likely to report using their medication for more than 30 days (41.6% vs. 34%, p5.03). Days of medication use (1–30 vs. 31+) among those using the medication was

S23

Table 1. Demographics, tobacco use, and quitting variables by intervention group for subjects who completed a 3-month follow-up survey (percent). Characteristic Demographics Age 50 years or older Female Some college or more Married Tobacco use Smoked less than one pack/day Years smoked (SD) Quitting behaviors Stayed quit for 24 hours Used medication.30 days* Used the quit aid Medication used Bupropion (Zyban) NRT Bupropion+NRT Used support from the manufacturera

Intervention Control group (n5515) group (n5520) 43.7 58.8 68.0 70.3

45.6 56.3 67.8 65.6

34.3

31.5

24.9 (12.7)

25.8 (12.6)

74.3 41.6 86.5

72.6 34.0 85.7

61.2 23.5 15.4 3.7

61.7 24.3 14.0 2.5

Note. aSurvey respondents were asked to select from a list of support actions they had used to quit or try to quit smoking. Included were responses about using phone or Internet support from the company that makes Zyban or the companies that make nicotine patch, gum, or inhaler. *p,.05.

not related to demographic variables, tobacco use, or chronic disease status. In addition, members were asked if they had used telephone or Internet support services available from the manufacturers of bupropion (Zyban) or NRT. Only 3.1% reported using these services. Quit attempts The majority (73.4%) of survey participants reported making one or more quit attempts lasting at least 24 hours. Fewer years of smoking (1–10 years) was associated with an increased likelihood of reporting a quit attempt, compared with a longer (11 or more years) smoking history (83.8% vs. 71.6%, p5003). Participants who reported using the medication were more likely than nonusers to make a quit attempt (78.8 vs. 38.8%, p,.001). No differences by intervention or chronic disease were observed. In addition, no demographic variables were associated with making a quit attempt. Self-reported quit rates At the 3-month follow-up, members randomized to the intervention group reported a 5.7% increase in point-prevalent quitting beyond the control group (Table 2). Smokers using less than one pack per day were 1.45 times more likely to quit, compared with those smoking one pack per day or more. A large and significant effect was observed for the duration

S24


of medication used compared with respondents who reported no use of the medication. No differences by demographics or chronic disease in quitting smoking were observed. In addition, type of medication used or years of smoking were not associated with pointprevalent quitting. In an additional analysis, we examined the quit rates by whether or not CHP made contact with a member. Members who were reached were significantly more likely to report quitting, compared with those who were not reached or not in the intervention group (37% vs. 28%, p5.005). Predictors of 3-month outcome We constructed multivariate logistic regression models of quitting smoking (Table 3). As a first step we determined that the chronic disease–6–treatment condition interaction was not significant (p5.09). However, significant interactions were observed for treatment condition by gender and by amount smoked at the time of the prescription fill. Therefore, we treated the effects of gender and amount smoked as stratification variables in Table 3. To further illustrate the effect of the intervention, Figure 2 summarizes model-predicted quitting rates Table 2. Associations with 3-month point-prevalent smoking cessation.

Treatment group* Intervention Control Age ,50 years 50+ years Gender Female Male Education Less than college Some college or more Marital status Married Other Chronic disease Yes No Amount smoked before medication use** One pack/day or more Less than one pack/day Years smoked ,11 11–30 31+ Duration of medication used*** 0 days 1–30 days 31+ days *p,.05; **p,.01; ***p,.001.

7-Day quit (percent)

Univariate odds ratio (95% confidence interval)

33.1 27.4

1.31 (1.01–1.71) Reference

27.8 33.3

Reference 1.30 (0.99–1.69)

30.5 29.9

Reference 0.97 (0.74–1.27)

26.7 31.9

Reference 1.28 (0.96–1.72)

30.8 29.1

1.09 (0.82–1.43) Reference

29.8 30.7

0.96 (0.74–1.25) Reference

27.7 35.8

Reference 1.45 (1.10–1.93)

35.4 28.2 31.0

Reference 0.72 (0.49–1.06) 0.82 (0.54–1.25)

6.3 23.9 52.7

Reference 4.65 (2.30–9.40) 16.46 (8.09–33.48)

by treatment group stratified by gender and amount smoked. These rates are based on the model shown in Table 3 and evaluated for individuals aged 50 years or older with one or more chronic conditions who used the smoking cessation medication for 1–30 days. Among women who smoked one pack or more a day, those in the proactive contact group were more likely to quit than those in the comparison group. However, for males smoking one pack or more a day, the intervention had no effect. In contrast, males who smoked less than one pack a day had a lower quitting rate if they were in the intervention condition. However, for females smoking less than one pack a day, the intervention had no effect. Beyond the effects of the intervention, we found that smokers aged 50 years or older were significantly more likely than younger smokers to quit. In addition, the odds of being a quitter were 16 times greater among those who used more than 30 days of medication, compared with those who did not use the medication.

Discussion Providing proactive telephone counseling as an adjunct for smoking cessation medication had an Table 3. Multiple logistic regression model predicting 3-month point-prevalent smoking cessation. Odds ratio (95% confidence interval) Treatment group6gender interaction Treatment group6baseline smoking interaction a Treatment group (effect for women who smoked one pack/day or more at baseline)b Not invited Invited Age ,50 years 50+ years Chronic disease No Yes Duration of medication used 0 days 1–30 days 31+ days

a

0.43 (0.23–0.80) 0.44 (0.24–0.83)

Referent 2.44 (1.51–3.94) Reference 1.49 (1.07–2.07) Reference 0.77 (0.56–1.07) Reference 4.21 (2.07–8.58) 15.62 (7.62–32.00)

Note. aGender (reference category: female) and baseline smoking (reference category: smoked one pack/day or more) are included in the model and in two-way interactions with treatment group. Model parameters for gender and amount smoked are not listed because these items are viewed as stratifying variables in the interaction terms with treatment group. b The model parameters for the treatment group effect for women who smoked less than one pack/day at baseline are OR51.08 (2.4460.44), 95% CI5.62–1.87. The model parameters for the treatment group effect for men who smoked less than one pack/ day at baseline are OR50.47 (2.4460.4360.44), 95% CI50.25–0.89. The model parameters for the treatment group effect for men who smoked one pack/day or more at baseline are OR51.05 (2.4460.43), 95% CI50.64–1.73.


effect on point-prevalent quitting 3 months later, but this effect varied by gender and amount smoked at the time the prescription was filled. In contrast to the positive effect of the treatment group on quit rates among females smoking more than one pack a day, male smokers who smoked less than one pack a day were less likely to quit if they were assigned to the treatment group (see Figure 2). The lower quit rate for males was found for those in the treatment group regardless of whether CHP was able to reach them to offer telephone counseling. This finding is reassuring because it suggests that being reached by proactive contact was not responsible for the lower quit rates among those in the treatment condition. We also found no differences among males in course enrollment and general counseling. However, one possible explanation for the higher quit rates among males in the control condition could be medication use. Use of the medication among males in the control condition who smoked less than a pack a day was higher (96.1%) than in any other group (medication use rates ranged from 78.7% to 88.8%). Other researchers have considered gender differences in smoking cessation and suggested that men are more likely to quit after using pharmacotherapy, compared with women (Perkins, 2001; Scharf & Shiffman, 2004). However, this finding could be a result of reduced success among women across all smoking cessation treatments (Fiore et al., 2000). In a recent meta-analysis, Cepeda-Benito, Reynoso, and Erath (2004) examined gender differences in the

Figure 2.

Model-predicted quit rates by treatment group.

S25

efficacy of NRT. For part of their analysis they rated studies as low or high on adjunct support, where high support involved a minimum of 30 minutes or two additional assessments. NRT given with a low level of supporting therapy increased quitting for men across all follow-up periods. However, NRT was efficacious for women regardless of the level of support in the short term, but over longer follow-up women using NRT experienced increased quit rates with more supporting therapy. Although the results of the present study differ in that supporting therapy increased quit rates among women in the short term, we conclude that nonpharmacological support may be more important for women than men. This conclusion suggests that physicians and clinics should offer more therapy to smokers and is congruent with recommendations for all smokers to use additional therapy instead of relying on pharmacotherapy to quit smoking. The present project adopted a population-based approach of using mail and telephone to contact participants. With no face-to-face contact, reaching members assigned to the treatment group was an increased challenge. For the purposes of evaluation in a health plan setting, we adopted the existing protocol used by CHP in other proactive outreach efforts. CHP was able to reach half of the members assigned to the treatment group using three attempts to reach the health plan members at their home telephone number. An important advantage of the present study is that it tested an intervention that,

S26


realistically, is likely to be used by other health plans. As a result, it has generalizable potential if successful beyond 3 months of follow-up. In addition to the task of reaching members by telephone, we faced the challenge of making an offer of telephone counseling close to the time the prescription was filled. CHP began calling members on average 7 days after the fill date. Given the importance of the first few days of not smoking on the success of a quit attempt, our delay in contacting individuals may have affected the uptake of counseling. Perhaps an offer of behavioral support at the time the prescription was filled would have generated more interest. However, among those we reached who declined to participate in additional telephone counseling, the counselor reported anecdotally a desire by patients to try the medication first before seeking help if they were not able to quit without it. Proactive telephone counseling as an adjunct to pharmacotherapy has been tested in a few studies. One of these studies examined the effect of using nursing staff to conduct three sessions of telephone counseling in addition to physician advice and NRT (Reid et al., 1999). The researchers found no difference in point-prevalent quit rates through 12 months of follow-up. An earlier study compared physician advice only to face-to-face counseling with and without nicotine gum (Ockene et al., 1991). In that study, face-to-face counseling plus nicotine gum produced the highest quit rates, but the addition of follow-up telephone counseling did not increase quit rates. In another study involving HealthPartners, no effect was found for the addition of telephone outreach with 509 smokers who were recruited with an offer of free nicotine patches (Lando et al., 1997). However, the impact of free patches plus proactive telephone peer support was effective in helping 214 low-income women quit smoking at 3 months, although the differences in favor of proactive calling were not sustained through 6 months of follow-up (Solomon et al., 2000). In another study with low-income smokers, Wadland and colleagues (2001) found telephone counseling effective at 3-month follow-up, compared with physician advice and medication. In a recent study, increased smoking cessation was reported through 6 months with 854 smokers randomized to NRT plus telephone counseling or NRT as transdermal patch, using a relapse-sensitive call schedule. In this trial, patients were given 2 weeks of nicotine patches for free and they could purchase additional weeks. In comparison, our members were required to obtain a prescription, but their medication was free with a nominal copayment. Interestingly, the quit rates reported in this study through 3 months (31.6% vs. 25.1%) were comparable with the observed self-reported quit rates (33.1% vs. 27.4%; Macleod et al., 2003).

Using more days of the medication was a significant factor in predicting quitting in the present study. Although the intervention group was more likely to use more days of medication, remaining abstinent may have reinforced continued use of the medication. Also, some other unmeasured variables may have enhanced continued use of the medication. The proportion of survey respondents who reported not using the medication (14%) is similar to data reported recently in a cohort study of German smokers identified after purchasing over-the-counter NRT (16.6% never used; Hasford, Fagerstro¨m, & Haustein, 2003). As a population-based intervention, the present study differed from clinic-based randomized trials in several ways, including the absence of biochemical verification of abstinence and lack of baseline measurements. Given the exclusive use of the telephone and mail for interaction, the demand on participants was low. This factor helps to offset the need for biochemical verification (SRNT Subcommittee on Biochemical Verification, 2002). In summary, we observed a short-term increased quit rate among health plan members offered telephone-based smoking cessation counseling as an adjunct to their filling of cessation pharmacotherapy. The success of the intervention varied by gender and amount smoked, with females in the intervention group smoking more than one pack a day more likely to quit. In addition, the characteristics associated with quitting included older age and use of more than 30 days of medication. Modifications in approach to the intervention (such as providing an offer of telephone support at the time a prescription is filled or follow-up contact to offer additional support) might provide stronger effects and should be studied.

Acknowledgments The authors would like to acknowledge Colleen King and the HPRF Data Collection Center for their work, Winnie Nelson for providing feedback on the paper, and Christopher Enstad for providing project management. This research was supported by grant 044163 from The Robert Wood Johnson Foundation’s Addressing Tobacco in Managed Care program.

References Cepeda-Benito, A., Reynoso, J. T., & Erath, S. (2004). Meta-analysis of the efficacy of nicotine replacement therapy for smoking cessation: Differences between men and women. Journal of Consulting and Clinical Psychology, 72, 712–722. Dillman, D. A. (2000). Mail and Internet surveys. The tailored design method. New York: Wiley. Fiore, M. C., Bailey, W. C., Cohen, S. J., Dorfman, S. F., Gritz, E. R., Heyman, R. B., Holbrook, J., Jaen, C. R., Kottke, T. E., Lando, H. A., Mecklenbur, R., Mullen, P. D., Nett, L. M., Robinson, L., Stitzer, M., Tommasello, A. C., Villejo, L., & Wewers, M. E. (2000). Treating tobacco use and dependence. Clinical practice guideline. Rockville, MD: U.S. Department of Health and Human Services. Public Health Service.

NICOTINE & TOBACCO RESEARCH Hasford, J., Fagerstro¨m, K. O., & Haustein, K. O. (2003). A naturalistic cohort study on effectiveness, safety and usage pattern of an over-the-counter nicotine patch. Cohort study on smoking cessation. European Journal of Clinical Pharmacology, 59, 443–447. Jaen, C. R., Crabtree, B. F., Zyzanski, S. J., Goodwin, M. A., & Stange, K. C. (1998). Making time for tobacco cessation counseling. Journal of Family Practice, 46, 425–428. Jaen, C. R., Stange, K. C., Tumiel, L. M., & Nutting, P. (1997). Missed opportunities for prevention: Smoking cessation counseling and the competing demands of practice. Journal of Family Practice, 45, 348–354. Lando, H. A., Rolnick, S., Klevan, D., Roski, J., Cherney, L., & Lauger, G. (1997). Telephone support as an adjunct to transdermal nicotine in smoking cessation. American Journal of Public Health, 87, 1670–1674. Lichtenstein, E., Glasgow, R. E., Lando, H. A., Ossip-Klein, D. J., & Boles, S. M. (1996). Telephone counseling for smoking cessation: Rationales and meta-analytic review of evidence. Health Education Research, 11, 243–257. MacLeod, Z. R., Charles, M. A., Arnaldi, V. C., & Adams, I. M. (2003). Telephone counseling as an adjunct to nicotine patches in smoking cessation: A randomised controlled trial. Medical Journal of Australia, 179, 349–352. Ockene, J. K., Kristeller, J., Goldberg, R., Amick, T. L., Pekow, P. S., & Hosmer, D. (1991). Increasing the efficacy of physician-delivered smoking interventions: A randomized clinical trial. Journal of General Internal Medicine, 6, 1–8. Ossip-Klein, D. J., & McIntosh, S. (2003). Quitlines in North America: Evidence base and applications. The American Journal of the Medical Sciences, 26, 201–205.

S27

Ossip-Klein, D. J., McIntosh, S., Utman, C., Burton, K., Spada, J., & Guido, J. (2000). Smokers ages 50+: Who gets physician advice to quit? Preventive Medicine, 31, 364–369. Perkins, K. (2001). Smoking cessation in women. Special considerations. CNS Drugs, 15, 391–411. Reid, R. D., Pipe, A., & Dafoe, W. A. (1999). Is telephone counseling a useful addition to physician advice and nicotine replacement therapy in helping patients to stop smoking? A randomized controlled trial. Canadian Medical Association Journal, 160, 1577–1581. Scharf, D., & Shiffman, S. (2004). Are there gender differences in smoking cessation, with and without bupropion? Pooled- and metaanalyses of clinical trials of bupropion SR. Addiction, 99, 1462–1469. Simon, J. A., Carmody, T. P., Hudes, E. S., Snyder, E., & Murray, J. (2003). Intensive smoking cessation counseling versus minimal counseling among hospitalized smokers treated with transdermal nicotine replacement: A randomized trial. American Journal of Medicine, 114, 555–562. Solomon, L. J., Scharoun, G. M., Flynn, B. S., Secker-Walker, R. H., & Sepinwall, D. (2000). Free nicotine patches plus proactive telephone peer support to help low-income women stop smoking. Preventive Medicine, 31, 68–74. SRNT Subcommittee on Biochemical Verification. (2003). Biochemical verification of tobacco use and cessation. Nicotine & Tobacco Research, 4, 149–159. Thorndike, A. N., Rigotti, N. A., Stafford, R. S., & Singer, D. E. (1998). National patterns in the treatment of smokers by physicians. The Journal of the American Medical Association, 279, 604–608. Wadland, W. C., Soffelmayr, B., & Ives, K. (2001). Enhancing smoking cessation of low-income smokers in managed care. Journal of Family Practice, 50, 138–144.