We examined racial/ethnic variations in the use of nicotine replacement therapy (NRT) and quit ratios among Caucasian, African American, Asian, and Latino lifetime smokers ages 25 to 44 years. We conducted cross-sectional analyses using data from individuals (n = 27,031) screened for enrollment in the Collaborative Study of the Genetics of Nicotine Dependence. Participants were randomly sampled from three Midwestern metropolitan areas using Health Maintenance Organization membership lists in Detroit, MI and Minneapolis, MN and a driver’s license registry in St. Louis, MO from March 2003 to August 2005. A telephone survey collected information on smoking history, previous quit attempts, and sociodemographic characteristics. Among lifetime smokers (n = 9,216), univariate analysis indicated that African Americans (22%) and Latinos (22%) were significantly less likely to report having ever used NRT for smoking cessation than Caucasians (31%). Asians (22%) also reported lower rates of using NRT than Caucasians, but this difference was marginally significant (P = 0.06). These disparities persisted in multivariate analysis for African Americans [adjusted odds ratio (OR), 0.76; 95% confidence interval (95% CI), 0.63-0.91; P < 0.01] but not for Latinos (adjusted OR, 0.76; 95% CI, 0.54-1.06; P = 0.11) or Asians (adjusted OR, 0.98; 95% CI, 0.60-1.60; P = 0.95). As measured by the quit ratio, African Americans (35%) were less likely to have quit smoking than Caucasians (52%). This disparity persisted in multivariate logistic regression (adjusted OR, 0.66; 95% CI, 0.56-0.78; P < 0.001). Asian and Latino smokers were as likely as Caucasians to report smoking cessation. Future prospective studies are needed to assess whether lower utilization of cessation treatments such as NRT contribute to the observed disparity in quit ratios for African Americans. (Cancer Epidemiol Biomarkers Prev 2008;17(7):1640–7)
Increasing evidence indicates the presence of racial/ethnic disparities in tobacco use cessation rates (1, 2). For example, Latino, Asian, and African American smokers are more likely to attempt cessation than Caucasian smokers, but they do not report having higher successful cessation rates (1). In 2000, smoking cessation rates as indicated by the quit ratio (percentage of lifetime smokers who have quit smoking) were lower among African Americans (37.5%) and Hispanics (42.9%) compared with Caucasians (50.4%; ref. 2). These lower cessation rates do not appear to be explained by socioeconomic factors (3). Potential explanations for these racial/ethnic differences include, among others, different patterns of smoking behavior (e.g., differences in nicotine dependence) and utilization of evidence based cessation treatments, such as nicotine replacement therapy (NRT; refs. 4, 5). There is, however, a lack of information regarding the methods that different racial/ethnic minority groups use to stop smoking, especially among younger smokers.
Smoking cessation conveys tremendous health and economic benefits and the benefit yielded is greater for those who quit at a younger age. It is estimated that most of the excess mortality from smoking could be avoided by quitting by age 35 years (6). In the United States, smoking rates peak in the 25- to 44-year age bracket. In 2005, the current smoking rate was 24.1% among adults ages 25 to 44 years compared with 20.9% overall, a rate far higher than the Healthy People 2010 objective of <12% (7). Current national smoking cessation guidelines recommend NRT as a first-line therapy to increase the likelihood of successful cessation during a quit attempt (8). The nicotine patch, nicotine gum, and nicotine lozenge are available over-the-counter without a prescription. Despite the increased availability of NRT and other cessation treatments, trends in the quit ratio among adults ages 25 to 44 years stopped increasing in the 1990s and actually decreased from 37.1% in 1992 to 34.8% in 2000 (2).
There is also growing evidence of racial/ethnic disparities in tobacco treatment, including lower utilization of evidence-based treatment such as NRT (4, 5, 9, 10), lower rates of physician advice, and physician assistance to quit smoking (11-13). The first study to document disparities in NRT was an analysis of the 1996 California Tobacco Survey (9). African Americans, Latinos, and Asians were two to three times less likely to use NRT than Caucasians, but there was no adjustment for potential confounding from differences in sociodemographic factors and smoking history. Two subsequent studies verified disparities in NRT use adjusting for sociodemographic factors and smoking history, but one study only compared Latinos to Caucasians (5). Although it included a diverse sample of veterans, the other study was conducted primarily among male older smokers ages >50 years (4). No study has been conducted among a diverse sample of younger smokers. This represents a significant gap in the literature, especially given the importance of targeting younger smokers for smoking cessation interventions.
The purpose of the present cross-sectional analysis was to examine racial/ethnic variations in the utilization of NRT and quit ratios among a diverse, population-based sample of younger lifetime smokers ages 25 to 44 years. We examined the use of NRT because it is recommended by national smoking cessation guidelines for all smokers to aid quit attempts (8) and is the most common treatment used by smokers during aided cessation attempts (9). We tested the hypotheses that racial/ethnic minorities would be less likely to have ever used NRT for smoking cessation and less likely to have quit smoking as measured by the quit ratio. To address the limitations of prior studies that examined racial/ethnic differences, we conducted multivariate analyses controlling for racial/ethnic group differences in smoking history, including age of smoking initiation and nicotine dependence level, as well as sociodemographic characteristics.
Materials and Methods
Study Design and Sample
We conducted a cross-sectional analysis using baseline data from adults (n = 27,031) screened for enrollment in the Collaborative Study of the Genetics of Nicotine Dependence, a case-control study on the genetics of nicotine dependence. Individuals ages 25 to 44 years from three metropolitan areas in the Midwest were randomly sampled using Health Maintenance Organization membership lists in Detroit, MI and Minneapolis, MN and a driver’s license registry in St. Louis, MO. The response rate for the screening survey was 40% overall and similar across the three sites. For the present analysis, the sample was limited to lifetime smokers (individuals who had ever smoked >100 cigarettes) who were of Caucasian (n = 7,907), African American (n = 955), Latino (n = 246), and Asian (n = 108) race/ethnicity. Lifetime smokers who were multiracial or of other race (n = 387) were excluded.
Outcome Variables. The main outcome variables of interest for the present analyses were (a) prior use of NRT (e.g., the nicotine patch or gum) and (b) smoking status (as indicated by cigarette smoking in the past month).
Main Independent Variable. The main independent variable of interest was race/ethnicity, self-reported, and classified in the following categories: White/Caucasian, African American/Black, Hispanic/Latino, and Asian. Race and ethnicity were assessed as two separate questions on the telephone survey. Participants who indicated multiracial status or other race were coded as “other” and were not included in this analysis (n = 387). In this analysis, participants who indicated they were of Hispanic/Latino ethnicity were classified as Hispanic/Latino regardless of race.
Covariates. Covariates included age, gender, years of education, marital status (dichotomized as married or not married), and employment (categorized as full-time, out of work/disabled/retired, and part-time/homemaker/student). Smoking history and nicotine dependence were measured with several items, including the number of cigarettes per day at the time when participants were smoking the most, duration of time to first morning cigarette (categorized as within or greater than 30 min), age of initiation of regular smoking, and time of transition between first experimentation with cigarettes and regular smoking. Use of tobacco products, other than cigarettes and nicotine replacement products, such as cigars, was also measured.
Statistical Analysis. Presented first are the sample demographic and smoking history characteristics stratified by race/ethnicity. A series of logistic regression models were then estimated to examine the unadjusted effects of race/ethnicity, demographics, and smoking history on the use of NRT and on the likelihood of quitting smoking. To evaluate the independent contribution of race/ethnicity, two stepwise multivariate logistic regression models with backward elimination (criterion of P < 0.05) were used to model the links between race/ethnicity, demographic, and smoking history variables on the use of NRT and on the likelihood of quitting smoking. Only variables significantly associated with use of NRT or quitting smoking in univariate analyses were considered in the final backward selection models, with the exception of age and gender, which were forced into these multivariate models.
There were 27,031 participants who were screened for participation in the Collaborative Study of the Genetics of Nicotine Dependence case-control study. Of these, 9,603 (36%) were lifetime smokers and 4,764 (18%) were current smokers. The final sample included in this analysis consisted of 9,216 lifetime smokers and 86% were Caucasian, 10% African American, 3% Latino, and 1% Asian.
Almost all (97%) of the lifetime smokers ages 25 to 44 years had health insurance. The sample characteristics, stratified by participant race/ethnicity, are presented in Table 1. There were numerous demographic and smoking history differences by race/ethnicity. The average (SD) age of lifetime smokers was 36.4 (5.3) years and the majority were female, with the exception of Asian lifetime smokers who were mostly male. African Americans were less likely to be married and less likely to have advanced education. African Americans and Asians reported a significantly later age of initiation of regular smoking than Caucasians. Caucasians reported the heaviest levels of smoking compared with other participants; however, African Americans were more likely to report a shorter amount of time to their first cigarette of the day.
Use of NRT
Prior use of NRT was significantly associated with race/ethnicity, demographics, and smoking history in univariate analyses. For example, Caucasians were significantly more likely to report use of NRT than were African Americans and Latinos (see Table 2). Caucasians also reported higher rates of NRT use than Asians, but this difference was marginally significant (P = 0.06). Participants who were older, had ever used other tobacco products, and had higher levels of nicotine dependence when they were smoking the most, as indicated by an earlier time to first cigarette and heavier smoking, were also more likely to report use of NRT. To evaluate the independent effect of race/ethnicity, variables with significant univariate associations with quitting smoking were considered in a multivariate stepwise model (Table 3). All possible interactions for the selected variables were tested but were not significant at the 0.05 level. The final model indicated that African Americans [adjusted odds ratio (OR), 0.76; 95% confidence interval (95% CI), 0.63-0.91; P < 0.01] were less likely to have ever used NRT products for smoking cessation than Caucasians. The differences between Latinos and Asians compared with Caucasians observed in univariate analysis were no longer present. Both heavier smoking and an earlier time to first cigarette increased the odds of using NRT. Other independent factors associated with increased likelihood of use of NRT included female gender, having ever used other tobacco products, and higher levels of education.
Smoking Cessation/Quit Ratio
Among lifetime smokers, significant factors negatively associated with smoking cessation in univariate analysis were African American race/ethnicity, higher levels of nicotine dependence when they were smoking the most, and being out of work, disabled, or retired (see Table 4). Significant factors positively associated with smoking cessation were older age, being married, and having more years of education. The final adjusted multivariate logistic regression model showed that African Americans were significantly less likely to quit smoking than were Caucasians (OR, 0.66; 95% CI, 0.56-0.78; P < 0.001; Table 5). Asian and Latino lifetime smokers were as likely as Caucasians to report smoking cessation. Nicotine dependence level was negatively related to cessation. Participants whose first daily cigarette was smoked within 30 min of wakefulness were nearly half as likely to have quit smoking (OR, 0.55; 95% CI, 0.50-0.61; P < 0.001). Higher levels of cigarettes per day during a participant’s peak period of smoking were associated with reduced odds of cessation (OR, 0.81; 95% CI, 0.76-0.86; P < 0.001). In addition, higher levels of education, older age, being married, and male gender were independently associated with increased likelihood of cessation.
Data regarding racial/ethnic variations in use of NRT among smokers ages 25 to 44 years, a critical age for intervention, are lacking. In this sample of lifetime smokers ages 25 to 44 years, African American smokers were significantly less likely than Caucasian smokers to have ever used NRT for smoking cessation. This disparity persisted even after controlling for sociodemographic factors and smoking history (e.g., nicotine dependence). Although Latino and Asian smokers were less likely to use NRT in unadjusted analyses, this finding did not persist after controlling for sociodemographic factors and smoking history.
Our findings provide further evidence of racial/ethnic disparities in tobacco treatment by being consistent with findings of previous population-based studies documenting disparities in the use of NRT (9, 10). Our findings are also consistent with a study of veterans receiving care from the Veterans Health Administration, an equal access health-care system (4). This study consisted primarily of male smokers ages >50 years and found that African American smokers were significantly less likely to use NRT than Caucasian smokers after adjusting for sociodemographic factors, smoking history, psychological factors, and health status (adjusted OR, 0.53; 95% CI, 0.34-0.83). Our study, however, did not replicate the findings of a study using data from the 2001 Colorado Tobacco Attitudes and Behaviors, which observed that Latino smokers were less likely than non-Latino smokers to use NRT (adjusted OR, 0.31; 95% CI, 0.17-0.57; ref. 5). In our study, the number of Latinos was relatively small and the point estimate of the effect was in the same direction but not statistically significant.
Compared with national samples (e.g., the National Health Interview Survey), our study sample consisted primarily of individuals with health insurance with lower than national average rates of lifetime and current smoking. This may limit generalizability of the study findings as individuals in this study likely have greater access to health care than national samples. Despite this, we still observed significant racial/ethnic disparities, suggesting that factors beyond access to care are responsible for the observed disparities in NRT utilization. Provider and system-level factors are likely sources. Although guidelines recommend that health-care providers address tobacco use at every clinical visit, there is often a lack of skill, interest, and time to deliver quality tobacco cessation interventions. Time constraints also limit actual delivery of cessation services during busy clinic visits and providers may be less likely to intervene with racial/ethnic minority smokers who may have more competing demands and “more pressing” medical concerns. In addition, physicians’ interpersonal behaviors have a profound effect on patient utilization, adherence, and outcomes (14). This is especially relevant for racial/ethnic minorities. Studies of disparities in quality of provider interpersonal behavior repeatedly show lower encounter quality when Caucasian physicians are interacting with non-Caucasian versus Caucasian patients (15). For example, low income and African American race are predictive of physicians adopting a “narrowly biomedical” communication pattern, characterized by low patient control of communication, high levels of physician biomedical information giving, and close-ended question-asking (16). Interventions addressing provider and system barriers to delivery of smoking cessation treatments for racial/ethnic minority populations are sorely needed.
Recent findings suggest that patient beliefs, attitudes, and preferences may also be important determinants of decisions to use smoking cessation treatment. Audrain-McGovern et al. (17) examined correlates of participation in a smoking cessation trial among younger adults (ages 18-30 years). Race was a significant predictor of participation in the trial and Caucasians, compared with non-Caucasians, were six times more likely to participate in the behavioral smoking cessation interventions. Thus, even when smoking cessation care is available, minority smokers may be less likely to use it. Potential reasons that may contribute to minority smokers’ decisions to not use treatment include mistrust of physicians, negative attitudes toward treatment, skepticism about their effectiveness, lack of knowledge regarding the functional benefits of treatment (e.g., medications can be used to relieve withdrawal symptoms), and concerns about medication side effects (18). More research is needed to understand the role of patient beliefs, attitudes, and preferences in decisions to use smoking cessation care and to identify effective strategies for increasing consumer demand for evidence-based cessation treatments.
Our study is also consistent with prior research that African American smokers are less likely to quit smoking (1, 2, 19, 20), independent of sociodemographic factors (3), and adds to this research because we also controlled for smoking history (e.g., nicotine dependence and age of initiation). However, other studies indicate that controlling for population differences in age of smoking initiation (21) or sociodemographic factors (22, 23) dramatically attenuates or eliminates observed differences between African Americans and Caucasians. For example, an analysis of the CARDIA study, a longitudinal study of young adults (18-35 years) observed markedly lower 10-year cessation rates among African Americans than Caucasians (25% versus 35%; crude OR, 0.62 for women; 19% versus 31%; crude OR, 0.52 for men), but these differences were no longer significant after adjustment for socioeconomic factors (22). A recent analysis of the National Health Interview Survey 1990 to 2000 found that in each year Caucasians were more likely to quit smoking than African Americans (23). The differences between Caucasians and African Americans diminished after 1994 and were substantially attenuated after controlling for sociodemographic factors, although still significant with adjusted OR near 1.5. It is possible that the racial/ethnic differences in our study findings are related to differences in sample size or in study design as we conducted a cross-sectional study, whereas the CARDIA study was longitudinal and the King et al. study analyzed multiple years of cross-sectional data.
Possible explanations for the observed differences in the smoking cessation quit ratio between African Americans and Caucasians include differences in smoking patterns. African Americans are more likely to smoke menthol cigarettes that are higher in tar and nicotine and hence may be more addictive and more difficult to quit smoking compared with plain cigarettes (24). In the United States, 69% of African Americans smoke menthol cigarettes compared with 23% of Caucasians, 29% of Latinos, and 29% of Asians (25). Further, menthol smoking rates are particularly high among younger smokers. It has been suggested that menthol cigarette smoking may contribute to the existing tobacco-related health disparities between African Americans and Caucasians (26). However, only a few studies have examined the association between menthol cigarette smoking and cessation (24, 27-29). The findings from these studies are inconclusive. A recent secondary analysis of a randomized controlled trial of bupropion among 600 African Americans found that menthol smoking was associated with lower short-term (6 weeks) cessation but not long-term (6 months) cessation (30). Of note, the decreased short-term cessation rates for menthol smokers was only observed for participants who were ages <50 years and who had received bupropion.
Another possible explanation for the observed racial/ethnic differences in quit ratios is differences in access and utilization of smoking cessation care. In this study, we observed significant racial/ethnic disparities in the use of NRT, which raises the question of whether lower rates of NRT use contribute to lower smoking cessation rates among African Americans. Our study, however, is not able to address this question. A clear temporal sequence cannot be established due to the cross-sectional nature of the study design. If we attempted to examine the association between prior use of NRT and quit ratios, we would not know whether we are examining the effect of using NRT on cessation versus the effect of continuing to smoke on likelihood of using NRT.
The strengths of this study include the size and diversity of the study sample, particularly for African Americans, as data regarding smoking cessation behaviors in racial/ethnic minority groups are sparse. However, the cell sizes for Latinos and Asians were relatively small and limit reaching firm conclusions about these latter two groups. Another strength of the study is that we controlled for smoking history (e.g., nicotine dependence level) as well as sociodemographic factors in our multivariate analysis. Our study, however, has several potential limitations. First, in regard to the use of race/ethnicity in this analysis, the operational definition allows racial comparisons (e.g., African American versus White) to be disentangled from ethnicity. However, because Latino ethnicity is not exclusive to one particular race, for comparisons between Caucasians and Latinos, one is not able to separate out the independent contribution of Hispanic/Latino ethnicity from race. Second, we relied on self-reported assessment of use of NRT and smoking status, but previous studies indicate that misreporting rates are low in minimal intensity studies such as those involving surveys (31). Third, the overall response rate for the screening survey was 40% and there is potential for nonresponse bias. However, we would expect nonrespondents to be more disadvantaged than respondents, and if response bias was present, it would likely have a null effect on our findings. In other words, if bias is present, the observed racial/ethnic differences may be an underestimate.
An interesting observation was that education and use of NRT was inversely associated in unadjusted analyses but positively associated in multivariate analysis. We suspect that this association between education and NRT use in unadjusted analyses is confounded by nicotine dependence level. Education and nicotine dependence are strongly associated in that individuals with lower education have higher nicotine dependence levels (higher cigarettes per day and shorter time to first cigarette). After adjusting for nicotine dependence, education is positively associated with use of NRT. In previous research, higher education has been observed to be associated with greater likelihood of using NRT (32).
In conclusion, further research is needed to assess the potential factors contributing to the observed racial/ethnic disparities in utilization of NRT and smoking cessation quit ratios. For example, research is needed examining the patient, provider, and system-level factors that contribute to racial/ethnic disparities in use of NRT. Future prospective studies are needed to assess whether lower utilization of cessation treatments such as NRT contribute to the observed disparity in quit ratios for African Americans. In addition, the role of menthol cigarette smoking and its effect on cessation should be examined in prospective studies. Environmental, social, cultural, and genetic factors may also be important. Current national smoking cessation guidelines recommend that all smokers be offered tobacco cessation treatment that includes pharmacotherapy. In particular, the 25- to 44-year age group should be a priority target population for cessation interventions because this age group (a) has high rates of smoking and (b) stands to yield the greatest health benefits. However, due to being relatively young and healthy, this age group may have less frequent health-care provider visits where most cessation interventions may occur. Therefore, alternative approaches for tobacco cessation intervention outside the health-care setting should be considered. Finally, culturally appropriate interventions need to be developed to increase use of evidence-based pharmacologic and behavioral treatments during attempts to quit smoking by racial/ethnic minority smokers.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant support: Veterans Affairs Health Services Research and Development research career development award (S.S. Fu); and Veterans Affairs Health Services Research and Development postdoctoral research fellowship (M.M. Kodl); National Cancer Institute P01 CA089392 and University of Minnesota Transdisciplinary Tobacco Use Research Center P50DA13333.
Note: Portions of this article were presented as an oral abstract at the 2007 Society for Research on Nicotine & Tobacco Annual Meeting in Austin, Texas.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the federal government.