Objective: To compare smoking-cessation interest and behaviors in younger and older smokers to develop better smoking-cessation strategies for younger smokers.
Study Design: Mail survey with phone follow-up of age-stratified members of a large midwestern health plan to identify current and former smokers, followed by a second follow-up survey of this subgroup 12 months later.
Methods: The follow-up survey asked about change in smoking status, quit attempts, interest in quitting, and experience with various aids to quitting. Analyses compared adults age 18-24 years with adults age 25-65 years.
Results: Follow-up surveys were completed by 66.5% of subjects. Young adults smoked at much higher rates than older adults (24.5% vs 17.1%), but were less likely to smoke daily or to smoke as many cigarettes. Young adults were as likely to be interested in quitting and more likely to report a quit attempt in the past year (60.6% vs 49.6%; P = .009), but these attempts were much more likely to be unaided (51.2% vs 33.7%; P = .0003). They also were more likely to report decreasing smoking in response to new restrictions on smoking in restaurants and bars (37.2 % vs 24.7%; P = .001).
Conclusions: Higher rates of smoking among young adults donâ€™t reflect less interest in quitting, fewer quit attempts, or less success in quitting compared with older adults. However, their reports of receiving or using much less help in quitting suggest that health plans and clinicians might be able to increase this groupâ€™s cessation with more active support.
(Am J Manag Care. 2007;13:626-632)
This project was conducted at HealthPartners, an integrated healthcare system providing medical services through a large network of owned and contracted clinics. As one of the largest healthcare organizations in Minnesota, HealthPartners provides coverage to more than 670 000 members. About 30% of these enrollees receive care in 1 of the 22 multispecialty owned clinics (HealthPartners Medical Group).
Identification of Smokers
Baseline survey items were adapted from a recent survey of health plan members,5 Minnesota state surveys,6 and national surveillance system surveys.7 The survey instrument included information about current health status (rating of health, any diagnosis of asthma or depression), smoking status, and demographic characteristics (sex, race, employment, school enrollment, education, health insurance type). Current and infrequent smokers were asked about previous quit attempts, use of medications to quit, and interest in quitting.
To increase the response rate, we limited the baseline survey to 15 questions over 2 pages. We conducted a cognitive pretest of the instrument with health plan members (n = 16) before administration to the entire sample, and made minor changes. The follow-up survey used the same smoker categories, but items were added to assess change in smoking status between baseline and follow-up. The follow-up survey also included questions about use of cessation services. The format and content of follow-up questions were based on information gathered from focus group members (n = 7), who also completed a pilot version of the instrument.
Data Collection Procedures
Contingency tables and Pearson ?2 tests were used to compare the younger (age 18-24 years) and older (age 25-65 years) subjects. We initially compared 3 age groups: 18-24, 25-44, and 45-65, but the latter 2 groups were so similar that reporting on all 3 groups only served to confuse the comparisons.
Different subsamples of the data were utilized for different analyses. The sample of all baseline survey completers was utilized to describe the sociodemographic characteristics of survey respondents and to estimate smoking prevalence. The subsample of baseline survey completers who were current smokers was used to describe respondents with respect to sociodemographic characteristics, smoking frequency, quit interest, and quit attempts. Respondents who completed the 1-year follow-up survey and who also were former smokers, nondaily smokers, or daily smokers at baseline comprised the denominator in an analysis describing changes in smoking status over time. Respondents who completed the follow-up survey and who were daily smokers at baseline served as the denominator for analyses describing smoking behavior and quit interest at the follow-up as well as quit activities undertaken in the 12 months before the follow-up.
This study was powered at 90% to detect a 3.2% difference in current smoker rates when the rate in 1 group was 20% (20.0% vs 23.2%), and to detect a 2.8% difference when the rate in 1 group was 15% (15.0% vs 17.8%). These calculations assume use of the baseline survey completer sample (n = 7161), a 2-sided test of proportions, and a type I error rate of 0.05.
Baseline surveys were completed by 3756 younger adults and 3405 older adults. Response rates adjusted for 209 undeliverable surveys and 24 ineligible subjects (because of language barriers or death) were 69.1% and 78.6%, respectively. Among older adults, 88.5% of baseline surveys were completed by mail and 11.5% by phone (compared with 85.6% and 14.4%, respectively, for younger adults). The eligible sample for the follow-up survey consisted of 1352 younger adults and 1583 older adults who had indicated being current, infrequent, or former smokers. Follow-up surveys were obtained from 809 young adults and 1144 older adults for response rates of 60% and 72%, after adjusting for the same reasons as above. Older and younger adults who were daily or nondaily smokers at baseline had similar follow-up response rates (unadjusted rate of 62.0% for older adults, 58.1% for younger adults). However, among those not smoking at baseline, follow-up response rates differed by age group (unadjusted response rate of 78.4% for older adults and 63.6% for younger adults; P <.001). Among older adults, 89.2% of follow-up surveys were completed by mail and 10.8% by phone (compared with 75.2% and 24.8% for younger adults).
Among the older adult sample, nonrespondents to the survey were more likely than respondents to be male (54.1% vs 44.4%; P <.001), to be younger (mean age 39.8 vs 44.1 years; P <.001), and to have insurance through Medicaid (9.0% vs 3.2%; P <.001). There were no differences with respect to metro/nonmetro residence. Among the younger adult sample, nonrespondents to the survey were more likely than respondents to be male (50.8% vs 43.5%; P <.001) and to have insurance through Medicaid (15.9% vs 10.5%; P <.001). There were no differences with respect to age or metro/nonmetro residence.
Attributes of Baseline Survey Respondents
In Table 1 we report the characteristics of the current smokers in the baseline survey (ie, the infrequent smokers and quitters were eliminated) separately for daily and nondaily smokers. Among these current smokers, daily smoking was more common among the older than the younger smokers (72.3% vs 60.6%; P <.001). The sociodemographic differences between the younger and older adult samples noted above among all respondents were generally true of this subset of current smokers, although some differences occurred only for the daily smokers. For example, among daily smokers only, younger smokers were less likely than older smokers to smoke more than 10 cigarettes a day, to have a cigarette within 5 minutes of waking, and to be interested in quitting in the next 6 months. However, among the nondaily smokers, younger and older smokers were equally likely to be interested in quitting in the next 6 months. Younger and older smokers were equally likely to report making a quit attempt in the past year. Nearly all of the current smokers had close friends who smoked, but for young adult smokers, this was universal.
Change in Smoking Status from Baselineto Follow-up 1 Year Later We compared and tested for differences in the distributions of follow-up smoking status by age after first stratifying on baseline smoking status at baseline (Table 2). Table 2 also illustrates change in smoking status in the 12 months from baseline to follow-up. Those who were former and daily smokers at baseline showed differences in their follow-up smoking status by age group, with younger adults more likely than older adults to have changed smoking status between the baseline and the follow-up measurements. However, the most change between baseline and follow-up smoking status occurred among nondaily smokers, where 44% of those who were nondaily smokers at baseline migrated into other smoking categories at follow-up. For this group of nondaily smokers at baseline, follow-up smoking patterns were identical for both age groups.
To test whether quit rate differences (ie, not smoking at follow-up) by age group among daily smokers (11.4% for younger adults, 12.7% for older adults) existed after controlling for the higher cigarette smoking frequency and addiction found in the older adult group, we conducted a logistic regression analysis predicting nonsmoking status versus daily or nondaily smoking at follow-up. This model used only data from individuals who were daily smokers at baseline and included predictors of age group (younger/older), amount smoked at baseline (>10 cigarettes per day vs =10 cigarettes per day), and whether smoking occurred within 5 minutes of waking (yes/no). The age group association still was not significant (odds ratio = 0.65; 95% confidence interval = 0.37, 1.14) but continued in the direction observed in the bivariate analysis, with a lower quit rate among the younger compared with the older daily smokers. Computing follow-up quit rates for baseline daily smokers within strata defined by either the amount smoked at baseline or smoking within 5 minutes of waking yielded similar results, with quit rates always directionally lower among younger than older adults, but never statistically significantly different (data not shown).
Characteristics of Daily and Nondaily Baseline Smokers 1 Year Later Table 3 summarizes the smoking and quitting behaviors at follow-up of the 578 individuals identified as daily smokers and the 308 identified as nondaily smokers in the baseline survey, some of whom were not smoking at follow-up. Among daily smokers, young adults were more likely to report making quit attempts in the past year, but much less likely to use any resources to aid them in those quit attempts. They primarily had tried to quit cold turkey and infrequently reported using help from a doctor or medical clinic. Among nondaily smokers there were far fewer differences by age group, with younger smokers less likely to use icotine-replacement therapy or help from a doctor or medical clinic.
In the follow-up survey, we had asked daily smokers at baseline to report on clinician actions at their most recent doctor's office visit. Among those who reported any medical discussion of smoking (66.9% of young adults and 72.5% of older adults; P = .15), similar proportions of younger (98.1%) and older (98.4%) adults reported being asked whether they smoked. Younger and older smokers also were equally likely to report being asked to quit smoking (77.6% and 74.9%, respectively). But older adults were more likely to report being asked by a clinician to set a quit date than younger adults (19.3% vs 9.4%; P = .004), and marginally more likely to have a clinician suggest that they use a cessation medication (31.6% vs 23.4%; P = .07).
Impact of Local Smoking RestrictionsOn March 31, 2005 (about 4 months before our follow-up survey), smoking bans in restaurants and bars went into effect in several cities and counties in the metropolitan area, including Minneapolis and St. Paul. In the follow-up survey we asked current smokers whether the new smoking restrictions were causing them to smoke less or to try to quit. More than a third of younger adults reported smoking fewer cigarettes compared with about a quarter of older adults (37.2% vs 24.7%; P = .001), but younger adults were just as likely as older adults to report trying to quit because of the restrictions (16.2% and 13.5%, respectively).
This study provides previously unavailable information directly comparing the smoking-cessation patterns and behaviors over time of a broad sample of younger and older adult smokers. Compared with the older adults, the young adults were lighter, less frequent smokers who were still fluctuating in their smoking behavior over time. Contrary to popular belief, they were as interested in quitting, more likely to make quit attempts, and as likely to have quit 1 year later. However, their quit attempts were much more likely to be made without medication or counseling assistance, and they were much less likely to have received help from medical sources. (As plan members, these young adults had greater access to assistance and medications, so the finding that so few use smoking-cessation aids is even more striking.) New environmental restrictions on public smoking were at least as likely to affect their smoking as that of older adults, perhaps in part because this age group tends to spend more time in restaurants and bars.9
What makes these findings particularly important and unique is that these subjects were broadly representative of younger and older smokers in Minnesota. At least in Minnesota, where only 6% of the population is uninsured and most low-income people on medical assistance are included in health plan rolls through a contract with the state, an insured population reflects the community quite well. This is important because most of the existing literature on smoking and cessation behaviors of young adults comes from studies of the minority who are students in 4-year colleges or are in the military.10-15 In a separate analysis of the young adult responders in this study, we demonstrated that college students are more likely to smoke fewer than 11 cigarettes per day and to have quit smoking at follow-up and are much less likely to smoke daily and to smoke immediately on awakening.16 However, we confirmed in this more representative sample the 3 separate studies of DeBernardo et al, Hines, and Morrison et al in college students, who also found that cold turkey was the predominant method for quit attempts, and there was little interest in assistance.11,13,15 We also confirm the findings of Schoenborn and colleagues, who found based on 1997-1998 National Health Interview Survey data that 54.2% of all smokers age 18-24 years had tried to quit smoking in the year prior to the survey, compared with 43.5% of those age 25-44 years and 39.4% of those age 45-64 years.17
These findings have many implications for interventions in this high-risk age group. Although more adults age 18-24 years smoke compared with older adults, these younger smokers are just as ready to quit, and they might quit in larger numbers if given more encouragement and support. Physicians and medical care organizations should pay particular attention, because these young smokers reported receiving much less help than older smokers did. That may be part of the reason why those young smokers who tried quitting were much less likely to use medications or other support, but recent studies both confirm our results and suggest that young adults have negative attitudes toward traditional smoking cessation approaches.18,19 Physicians seem well placed to change this pattern if they implement practice systems that make it more likely that individualized support and assistance are provided, going well beyond the usual Ask and Advise sections of the Public Health Service clinical guideline Treating Tobacco Use and Dependence.20 Health plans also might develop programs to reach out to assist those young smokers interested in quitting instead of assuming that they have to be persuaded to quit.
We chose to focus on these comparisons between younger and older adult smokers because the much higher rate of smoking in young adults and their status as the smokers of the future require greater efforts to encourage and support their cessation interests. Moreover, we specifically did not conduct logistic regression analysis of cessation predictor variables because we believe that such statistical findings are of dubious value for targeting interventions and distract from the main usable conclusions. These findings instead highlight the comparability of young adult smoking behaviors and interests, and what health plans, medical clinics, and individual physicians need to know to take action.
This study's limitations include the fact that all of the respondents are health plan members who provided self-reports. The response rate to the follow-up survey was lower than hoped and lower than that of the original survey, probably because we have demonstrated that smokers are known to be less likely to respond to surveys about this habit.5 In addition, young adults had lower response rates to both the baseline and follow-up surveys, potentially leading to an exaggeration of differences between them and older smokers. However, significant differences in response rates to the follow- up survey by age group were found only among those who were not smoking at baseline. Thus, quit rates and other characteristics (eg, quit behaviors, medical assistance for quitting) computed at follow-up for those who were smoking at baseline are unlikely to be affected by differences in response rates to the follow-up survey.
These results should be useful to policymakers, health promotion planners, and those in medical settings who want to reduce smoking prevalence by paying greater attention to the age group who have the highest prevalence of smoking and who will form the smokers of the future.
Author Affiliations: From HealthPartners Research Foundation, Minneapolis, Minn.
Funding Source: Any public dissemination of information relating to the grant was made possible by Grant Number RC-2006-0011 from ClearWay MinnesotaSM. The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of ClearWay Minnesota.
Author Disclosures: The authors (LIS, RGB, MM, SEA, MJT) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (LIS, MCM); acquisition of data (LIS, MCM, MJT); analysis and interpretation of data (LIS, RGB, SEA, MJT); drafting of the manuscript (LIS, RGB, SEA, MJT); critical revision of the manuscript for important intellectual content (LIS, RGB, MCM, SEA); statistical analysis (SEA); obtaining funds (MCM); administrative, technical, or logistic support (MJT).
Address correspondence to: Leif I. Solberg, MD, HealthPartners Research Foundation, PO Box 1524, MS 21111R, Minneapolis, MN 55440-1524. E-mail: email@example.com. Johnston LD, Oâ€™Malley PM, Bachman JG. Monitoring the Future National Survey Results on Adolescent Drug Use: Overview of Key Findings, 2001. Bethesda, Md: National Institute on Drug Abuse; 2002. NIH publication 02-5105.
3. Cigarette smoking among adultsâ€”United States, 2001. MMWR. 2003;52:842-844.
5. Solberg LI, Hollis JA, Stevens VJ, Rigotti NA, Quinn VP, Aickin M. Does methodology affect the ability to monitor tobacco control activities? Implications for HEDIS and other performance measures. Prev Med. 2003;37:33-40.
7. National Center for Chronic Disease Prevention & Health Promotion. Behavioral Risk Factor Surveillance System. Trends data. Minnesotaâ€”grouped by age. Available at: http://apps.nccd.cdc.gov/brfss/Trends/agechart.asp?qkey=100000&state=MN. Accessed August 1, 2003.
9. Biener L, Albers AB. Young adults: vulnerable new targets of tobacco marketing. Am J Public Health. 2004;94:326-330.
11. DeBernardo RL, Aldinger CE, Dawood OR, Hanson RE, Lee SJ, Rinaldi SR. An e-mail assessment of undergraduatesâ€™ attitudes toward smoking. J Am Coll Health. 1999;48:61-66.
13. Hines D. Young smokersâ€™ attitudes about methods for quitting smoking: barriers and benefits to using assisted methods. Addict Behav. 1996;21:531-535.
15. Morrison K, Banas J, Burke M. Understanding college studentsâ€™ salient attitudes and beliefs about smoking: distinctions between smokers, nonsmokers, and ex-smokers. Public Health Rev. 2003;31:95-109.
17. Schoenborn CA, Vickerie JL, Barnes PM. Cigarette Smoking Behavior of Adults: United States, 1997-98. Hyattsville, Md: National Center for Health Statistics; February 7, 2003. Advance Data From Vital and Health Statistics, No. 331.
19. Curry SJ, Sporer AK, Pugach O, Campbell RT, Emery S. Use of tobacco cessation treatments among young adult smokers: 2005 National Health Interview Survey. Am J Public Health. 2007;97:1464-1469.