To our knowledge, this is the first study to examine differences among smokers enrolled in a cessation trial who reduced their level of smoking in comparison with those who quit smoking and those who did not reduce their smoking. Moreover, this research focuses on Black light smokers, a priority population given the higher rates of smoking, and their increased risk for smoking-related health problems. Thus, this paper contributes novel information about a critical public health issue.
First, in regard to smoking-related characteristics, we found that having lower baseline cotinine levels and lower baseline smoking rates were associated with being more likely to quit smoking than to reduce smoking, which is reasonable and in line with the findings of Nollen et al. (2006)
. However, higher nicotine dependence was associated with being quitter rather than a reducer, which seems contradictory and deserves further examination. Having higher baseline smoking levels predicted reduction in smoking rather than no reduction. One potential reason for this finding is that reduction may be easier to achieve if an individual smokes more at baseline versus if an individual smokes less. Despite reducers having higher baseline levels of smoking than nonreducers, they also had lower nicotine dependence and lower baseline cotinine levels than nonreducers. The reason for these somewhat contradictory findings is unclear. Perhaps the lower cotinine levels indicated depth of inhalation or use of menthol cigarettes among nonreducers, which have been shown to have an association with higher cotinine levels (Mustonen, Spencer, Hoskinson, Sachs, & Garvey, 2005
). Although using menthol cigarettes did not significantly contribute to the multivariate model, more nonreducers used menthol cigarettes than did reducers, as indicated by the bivariate analyses. One review (Ahijevych & Garrett, 2004
) indicated that menthol increases drug absorption and that smoking menthol cigarettes increases carbon monoxide levels at a greater rate than non–menthol cigarettes. Moreover, although there are conflicting reports on the effects of menthol on smoking topography, menthol may increase depth of inhalation, puff volume, and puff frequency, which has implications for cotinine levels and addiction (Ahijevych, Gillespie, Demirci, & Jagadeesh, 1996
; Djordjevic et al., 2002
). In addition, menthol smokers smoke their first cigarette earlier in the day than non–menthol smokers, indicating greater nicotine dependence (Ahijevych & Garrett). The current findings also suggest that nonreducers, despite smoking fewer cpd at baseline, may be more physiologically dependent (Chen, Blevins-Primeau, Dellinger, Muscat, & Lazarus, 2007
In terms of psychosocial variables, two major factors were significant predictors of cessation and lack of reduction in comparison with reduction—perceived stress and expected negative social impressions of smoking. In terms of stress, lower levels of stress were related to greater likelihood of not reducing smoking or quitting in comparison with reducing smoking, which is difficult to interpret. Prior work demonstrating that smoking cessation was associated with less negative affect (Ferguson et al., 2003
; K. J. Harris et al., 2004
) coupled with literature documenting distress tolerance as a predictor of ability to refrain from smoking (Brown, Lejuez, Kahler, & Strong, 2002
) supports the finding that less perceived stress was associated with successful cessation. However, lower baseline stress predicting no reduction versus reduction seems counterintuitive. Negative social impressions of smoking also did not demonstrate the anticipated trajectory; rather, those who were concerned about smoking-related negative social impressions were more likely to quit smoking or not reduce their smoking rather than reduce their smoking. Prior work has found that more negative expectancies and fewer positive expectancies related to smoking are related to greater readiness to quit smoking (Pulvers et al., 2004
). Thus, it is reasonable that greater expected negative social impressions of smoking predicting cessation versus reduction. Again, the fact that nonreducers had greater concern about negative social impressions in comparison with reducers is counterintuitive. Overall, how anticipated consequences of smoking and experience of stress may differentially predict reduction and cessation should be further examined.
In addition, higher BMI was associated with cessation versus reduction. Similarly, Nollen et al. (2006)
found that higher baseline BMI was related to quitting smoking. Perhaps concern about weight gain as a result of quitting smoking is less significant among those who have higher BMI. Weight gain concern has been found to predict continued smoking versus cessation (Meyers et al., 1997
) and attrition from smoking programs (Copeland, Martin, Geiselman, Rash, & Kendzor, 2006
). Further research is needed to understand the dynamics of BMI, weight concern, and smoking trajectories.
Another interesting finding was that those assigned to the MI arm of the randomized clinical trial were more likely to reduce their smoking rather than quit, but no relationship was found between reducers and nonreducers. It is unclear why this difference existed between reducers and quitters. The parent study (Ahluwalia et al., 2006
) found that HE was more effective in achieving cessation than MI, counter to the hypotheses of the study. One primary reason offered was that participants enrolled in the parent trial were already highly motivated to quit (scored 9/10). The early generation of MI provided in this trial was designed to increase intrinsic motivation but did little to provide problem solving guidance toward cessation. Therefore, the MI delivered may not have been sufficient to assist participants with the behavioral actions needed to achieve cessation. MI now encourages the provision of a behavioral action plan and the provision of direct assistance once motivation is achieved (Miller & Rose, 2009
). Thus, it may also be that HE is more effective for providing people who are motivated to quit with practical suggestions to aid in cessation.
In addition, our results regarding sociodemographics are consistent with some prior research. For example, women were more likely to reduce than to quit, which is in line with previous research documenting more successful cessation among men (Bjornson et al., 1995
; Wetter et al., 1999
). However, a significant relationship was not found when reducers were compared with nonreducers. This is in contrast to other research that found that men were more likely to reduce their smoking than women (Joseph et al., 2005
). Among our sample of Black light smokers, it appears that women were less successful at quitting smoking but not at reducing their smoking level. Also, being older was associated with a greater likelihood of quitting smoking rather than reducing smoking level.
Some important limitations to this study do exist. First, this sample of Blacks smoking an average of 10 cpd or less was drawn from a Midwest inner-city clinic. Thus, these findings may not generalize to all Black light smokers. Second, although biochemical verification was used to validate smoking status at baseline and follow-up, assessments of cpd smoked at baseline and Week 26 were based on self-report. This is an important issue because we defined a “reducer” as anyone who reduced their smoking by at least 50%. Thus, we cannot be certain that participants accurately reported their cigarettes smoked per day. Additionally, we excluded participants who reported cessation but were not cotinine verified. Thus, we cannot be certain about the outcome of those that we excluded. Related to this issue of cotinine verification, we did not assess cotinine levels at follow-up among participants who did not report cessation. Future research should include assessments of cotinine among all participants. Furthermore, the relatively large sample size and the comparisons among three groups in this study may have caused some spurious findings. Additionally, it is unclear why characteristics that were significant factors in the bivariate analyses did not contribute significantly to the multivariate model. It may be that some factors were related to one another, and thus, accounting for one variable in the model diminished the significance of another. Ultimately, this study is an initial step in understanding the unique nature of those who reduce their smoking. Future research may build from these findings.
The present study identified factors differentiating smoking reduction versus cessation and reduction versus no reduction. The impact of original smoking level and cotinine level deserve further examination, as our results indicate that they are not synonymous in terms of predicting reduction versus cessation versus no reduction. Expectancies, motivation, and confidence also differentially predict these different smoking behavior change patterns. Thus, attending to psychosocial variables is important in helping people reach their smoking goals whether they include reducing their smoking or quitting.