How Do Light and Intermittent Smokers Differ from Heavy Smokers in Young Adulthood: The Role of Smoking Restraint Strategies

Abstract

Light and intermittent smoking has become a prevalent pattern of use among young adults. Little is known about which factors differentiate light and intermittent smokers (LITS) from heavy smokers (HS) in young adulthood. In this study, we compare young adult LITS with HS with regard to demographic- and smoking-related variables, self-control abilities, and concrete strategies of smoking restraint. The data were collected as part of an Ecological Momentary Assessment (EMA) study with 137 German young adult smokers (M Age = 21.1 years, 46.0% female; 76 HS [≥10 cigarettes/day] and 61 LITS [≤5 cigarettes/day]). Participants were recruited over the Internet and completed a baseline questionnaire online. Several variables differentiated LITS and HS in a multiple logistic regression analysis: LITS reported fewer smoking friends (p < .001) and a higher self-efficacy to resist smoking (p < .01). Further, LITS smoking status was associated with reporting a past quit attempt (p < .05) and the use of smoking restraint strategies (counting, limiting, and purposefully not smoking cigarettes; p < .05). Notably, nicotine dependence and trait self-control abilities did not differentiate between LITS and HS. Our results point to the role of smoking restraint strategies and self-monitoring of smoking to limit the daily number of cigarettes smoked.

Light and intermittent smoking has become an increasingly prevalent pattern of tobacco use among young adults (Presson, Chassin, and Sherman 2002; Schane, Glantz, and Ling 2009). In Germany, recent data suggest that approximately one in three young adults is a current smoker and, among these smokers, about a third are heavy smokers (HS; >10 cigarettes per day [CPD]), another third are light smokers smoking up to 10 CPD, and the last third are nondaily—or intermittent—smokers (Pabst et al. 2010).

While often viewed as a transitional phase, light and intermittent smoking is not necessarily a gateway to heavier smoking (Caldeira et al. 2012; Levy, Biener, and Rigotti 2009). It is important to understand which factors are associated with light and intermittent compared to heavy smoking in young adult smokers who are at a relatively early stage in their smoking careers (White et al. 2009); such information may be useful for informing targeted interventions for this group of smokers.

Research on subtypes of adult smokers shows that HS differ from light and intermittent smokers (LITS) in several ways; for example, compared with HS, LITS have higher education and socioeconomic status (Wortley et al. 2003), higher self-control and lower impulsivity (Heyman and Gibb 2006), more autonomy over smoking (Wellman, DiFranza, and Wood 2006), and they are less dependent on nicotine (Shiffman et al. 2012). In addition, LITS have a higher likelihood to report quit attempts over the past year (Wortley et al. 2003). Previous research examining older (mean age >40 years) adult African American smokers further suggests that, compared to HS, LITS may consciously restrain their smoking by using various strategies such as limiting the number of cigarettes or smoking less than a complete cigarette (Okuyemi et al. 2002). However, little is known as to how the use of smoking restraint strategies differs between young adult LITS and HS.

The goal of this study was to compare young adult LITS with HS with regard to demographics, variables related to smoking and quitting, trait self-control abilities, and strategies of smoking restraint. Further, we examined whether potential differences in smoking restraint strategies between LITS and HS still held true after accounting for other between-group differences.

Methods

Overview

The current analyses are based on participants from a German Ecological Momentary Assessment (EMA) study over 4.5 months designed to monitor young adults' smoking behavior (Thrul, Bühler, and Ferguson 2014, 2015). The present article draws on data gathered from the baseline assessment. Study procedures were approved by the ethics commission of the German Psychological Society.

Procedure and participants

Participants were recruited through a Facebook advertisement campaign employing a recruitment strategy found to be effective in previous studies (Frandsen, Walters, and Ferguson 2014; Ramo et al. 2014). Ads were targeted at young adults 18–25 years old living in Germany, and additional recruitment postings were made in student Facebook groups. Supplementary recruitment strategies included in-class information sessions at universities and leaflets distributed at schools and companies based in Munich (e.g., BMW). To be eligible, applicants needed to be between 18 and 25 years old, possess a smartphone, report an intention to quit within the next six months, have smoked ≥100 cigarettes in their life, and have a smoking rate that fit with our pre-defined categories: LITS (≤5 CPD) or HS (≥10 CPD). The LITS category was selected in accordance with the previous literature (Shiffman and Paty 2006). Interest in quitting was included as a selection criterion in order to maximize the likelihood of observed changes in smoking during the EMA monitoring period, the primary objective of the larger study, as well as to increase the external validity of our findings for informing smoking cessation or reduction interventions for smokers who are at least somewhat ready to change. In total, 203 participants screened successfully and 155 completed the baseline assessment. Of these, 18 participants were excluded from the analyses, as they reported smoking patterns that did not comply with our pre-defined categories in the baseline assessment; this resulted in a final sample of 137 young adult smokers (M = 21.1 years, 46.0% female; 76 HS and 61 LITS).

Measures

Perceived smoking of others

Perceived smoking of mother and father (“Does your mother/father smoke?”) was assessed with two items (“yes” or “no”). Perceived smoking of friends was assessed with one item, and responses were recorded on a four-point scale (“all” to “no one”).

Nicotine dependence

Strength of nicotine dependence was assessed using the Hooked on Nicotine Checklist (HONC) (DiFranza et al. 2002). A sum score was calculated over all 10 HONC items.

Smoking behavior

Quantity and frequency of smoking in the last 30 days was assessed with one question each and used to form an index of cigarettes per day (CPD = [quantity × frequency]/30) (Kraus et al. 2013).

Smoking stability

The stability of the currently reported smoking behavior was assessed with one question (“For how long have you been smoking at your current smoking pattern?”), with responses dichotomized to 1–12 months and >12 months.

Quit motivation

Motivation to quit was assessed using a four-point scale (ranging from “very” to “not at all”).

Quit attempt

A quit attempt within the past six months was assessed using a single item (“yes” or “no”).

Self-efficacy to resist smoking

Self-efficacy to resist smoking in certain situations was measured with a 12-item measure designed for adolescents (Kremers, Mudde, and De Vries 2001). Responses were recorded on a five-point scale (“very easy” to “very difficult”). A scale was created by averaging over all 12 items and Cronbach's Alpha of the scale was .83.

Strategies of smoking restraint

Three items were used to assess strategies of smoking restraint. These items were selected from previous literature (Okuyemi et al. 2002; Scharf 2009) and reflected different strategies such as counting or limiting CPD, or purposefully forgoing smoking specific cigarettes (each measured using “yes”/“no” items). A sum-score was created to reflect the number of strategies used (0 to 3).

Trait self-control

Trait self-control was assessed with a 40-item measure that was developed for young adults (Gibbs, Giever, and Higgins 2003). The instrument assesses self-control abilities in general and in specific situations, and responses were recorded on a 10-point scale (“totally disagree” to “totally agree”). As in previous studies (Gibbs, Giever, and Higgins 2003; Gibbs, Giever, and Martin 1998; Muraven et al. 2005), the scale showed excellent internal consistency in our study (Cronbach's Alpha = .91).

Analytical strategy

Bivariate differences between LITS and HS were investigated using descriptive statistics and t- and Chi-square tests. Significant predictors of the bivariate analyses we subsequently included into a multiple logistic regression analysis predicting group membership (LITS vs. HS). All analyses were conducted with Stata 14 (StataCorp 2015).

Results

In bivariate analyses, compared to HS, more LITS were currently students and LITS reported having fewer parents smoking and friends smoking (Table 1). LITS reported a less stable smoking pattern, scored lower in nicotine dependence (HONC), had a higher quit motivation, and more LITS than HS reported a quit attempt in the previous six months. Lastly, LITS reported a slightly higher trait self-control than HS (p = .05), a higher self-efficacy to resist smoking, and LITS had a higher likelihood than HS to use smoking restraint strategies. Age and gender were not significantly associated with smoking status.

Table 1.

Sample characteristics of HS and LITS and bivariate tests of significance.

	Total (N = 137)	HS (N = 76)	LITS (N = 61)
Characteristics	N or mean (% or SD)	N or mean (% or SD)	N or mean (% or SD)
Demographics
Age	21.1 (2.1)	21.4 (2.1)	20.8 (2.0)
Gender (% female)	63 (46.0%)	34 (44.7%)	29 (47.5%)
Student	101 (73.7%)	47 (61.8%)	54 (88.5%)***
Smoking by others
Parental smoking	64 (46.7%)	42 (55.3%)	22 (36.1%)*
Friends smoking (all or most of them)	90 (65.7%)	60 (79.0%)	30 (49.2%)***
Smoking and quitting history
Daily smokers	78 (56.9%)	67 (88.2%)	11 (18.0%)***
Number of smoking days (out of past 30 days)	27.1 (4.4)	29.5 (2.0)	24.1 (4.8)***
Number of cigarettes per smoking day	11.2 (7.4)	16.5 (5.7)	4.5 (1.6)***
CPD	10.6 (7.6)	16.2 (5.7)	3.6 (1.2)***
Stability of current smoking (12 months or more)	85 (62.0%)	57 (75.0%)	28 (45.9%)***
HONC score	6.2 (2.7)	7.0 (2.7)	5.3 (2.5)***
Quit motivation	2.5 (0.7)	2.4 (0.8)	2.7 (0.7)**
Quit attempt (previous 6 months)	39 (28.5%)	15 (19.7%)	24 (39.3%)*
Self-control and smoking restraint
Trait self-control1	6.0 (1.1)	5.8 (1.3)	6.2 (0.9)2
Self-efficacy to resist smoking1	2.5 (0.7)	2.2 (0.6)	2.8 (0.6)***
Use of restraint strategies	1.36 (1.14)	1.03 (1.08)	1.77 (1.07)***

Note: HS = heavy smokers, LITS = light and intermittent smokers, df = degrees of freedom, CPD = cigarettes per day, HONC = Hooked on Nicotine Checklist; continuous variables were compared using t-tests, categorical variables using Chi²-tests;

¹based on n = 135 participants because two participants did not complete half or more items on the scale;

²significant at p = .05, calculated with adjustment of unequal variances between groups;

^*p < .05;

^**p < .01;

^***p < .001.

Significant variables from the bivariate analyses were subsequently tested in a multiple logistic regression analysis to predict smoking status (LITS vs. HS; Table 2). When including all predictors simultaneously, LITS smoking status was significantly associated with reporting fewer smoking friends, a past quit attempt, a higher self-efficacy to resist smoking, and the use of smoking restraint strategies (Table 2).

Table 2.

Multiple logistic regression examining characteristics of group membership (LITS vs. HS) (n = 135).

Characteristics	OR	95% CI
Student status	2.3	[0.6, 8.6]
Parental smoking	0.5	[0.2, 1.4]
Friends smoking (all or most of them)	0.1***	[0.0, 0.4]
Stability of current smoking (12 months or more)	0.6	[0.2, 1.6]
HONC score	0.8	[0.6, 1.0]
Quit motivation	1.5	[0.7, 3.0]
Quit attempt (previous 6 months)	5.4*	[1.5, 19.9]
Trait self-control	0.9	[0.6, 1.4]
Self-efficacy to resist smoking	3.7**	[1.4, 9.3]
Use of restraint strategies	1.8*	[1.1, 2.9]

Note: OR = Odds ratio, CI = Confidence interval, HONC = Hooked on Nicotine Checklist;

^*p < .05;

^**p < .01;

^***p < .001.

Discussion

Only a few previous studies have compared young adult LITS and HS, and our study is the first to investigate the use of smoking restraint strategies among these different subgroups in young adult smokers. First, we found that reporting fewer smoking friends was significantly associated with LITS status. Since young adults frequently smoke when they are with other smokers (Cronk and Piasecki 2010; Otsuki et al. 2008; Thrul, Bühler, and Ferguson 2014), smoking is influenced by perceived social norms (Etcheverry and Agnew 2008; Kobus 2003), and has been shown to cluster in social networks (Christakis and Fowler 2008), it is not surprising that the peer group of HS consists of more smokers. The different smoking patterns of LITS and HS may thus partially be explained by the presence of smokers in the peer group.

Our findings suggest that young adults' perceived self-efficacy to resist smoking also sets LITS apart from HS, with LITS reporting more confidence to resist smoking. These results are consistent with previous studies that have established resistance self-efficacy as a predictor of adolescent smoking acquisition and cessation/reduction (Kremers, Mudde, and De Vries 2001; Moan and Rise 2006; Thrul et al. 2013), and show that resistance self-efficacy is also an important construct with regard to smoking in young adulthood. The perceived self-efficacy to resist smoking may be an important underlying factor that enables young adults to restrain their smoking in the first place. Since young adulthood is the developmental phase in which a person may be exposed to numerous smokers in the peer group, it is important to equip young people with skills that allow them to resist smoking.

Two variables from the behavioral domain differentiated LITS from HS: a past quit attempt and the use of strategies of smoking restraint. Consistent with previous literature, we found that a quit attempt in the previous six months was associated with LITS smoking status (Wortley et al. 2003), which suggests that young adult LITS may be closer to smoking cessation than HS. Further we found that the use of strategies of smoking restraint was a strong correlate of LITS smoking status, even after accounting for other predictors in the multiple regression model. This result is consistent with what others have found in older adult LITS smokers (Okuyemi et al. 2002). Interestingly, trait self-control abilities did not differentiate between LITS and HS in the multivariate model. Given that a prior study found higher self-control among LITS (Heyman and Gibb 2006), and the current study also found this association in the bivariate analysis, it would be warranted to test in future studies if smoking restraint strategies mediate the relationship between self-control and LITS status.

Our findings suggest that it may be fruitful to teach strategies of smoking restraint to HS who want to limit and reduce their smoking. Useful methods of smoking reduction examined in the past include hierarchical or scheduled gradual reduction (Hughes and Carpenter 2005; Riggs, Hughes, and Pillitteri 2001; Riley et al. 2002) or nicotine replacement therapy (Carpenter et al. 2004; Hughes and Carpenter 2005). Training in self-control has been shown to lower the risk of smoking lapses and may be a useful concept to teach to smokers who want to reduce their smoking (Muraven 2010). An understanding of smoking restraint strategies that are successfully employed by LITS may be useful for HS who are not ready to quit but may be motivated to reduce their smoking. While smoking reduction may not necessarily result in substantially reduced harm (Schane, Ling, and Glantz 2010), it may be a precursor of successful quitting in the future (Hughes and Carpenter 2006), and thus a useful intermediate treatment goal.

Lastly, nicotine dependence did not differentiate between LITS and HS in the multivariate model, even though LITS reported smoking significantly fewer CPD than HS. LITS endorsed on average 5 out of 10 HONC items—this finding provides further evidence that LITS are also experiencing nicotine dependence, even though their levels may be lower than those found among HS (Shiffman et al. 2012). Taken together, our results suggest that, while LITS may experience symptoms of nicotine dependence, there are also protective factors in place that may help them deal with their dependence symptoms and smoking urges, and enable them to limit their daily number of cigarettes. These protective factors include a high self-efficacy to resist cigarettes, few smoking friends, a high readiness to quit smoking, as suggested by a higher likelihood of having tried to quit in the past six months, and the use of deliberate strategies of smoking restraint, such as counting and limiting their daily number of cigarettes, and purposefully forgoing smoking specific cigarettes.

Our results should be interpreted with limitations in mind. We used cross-sectional data and cannot draw conclusion on the directionality of our reported effects (e.g., self-efficacy to resist smoking may be both a predictor and a consequence of light and intermittent smoking behavior). Further, we used inclusion criteria to recruit specific groups of smokers who were ready to quit within the next six months, potentially limiting the generalizability of our results. However, more than 50% of young adult smokers in Germany report intentions to quit within the next 30 days to six months (Federal Centre for Health Education (BZgA) 2012), and our result may thus be generalizable to a substantial proportion of young adult smokers. Although we collected data on the stability of smoking patterns in the two groups, we do not know whether LITS were decreasing or increasing their smoking over time before participating in this study. While we included the observed smoking pattern stability as a covariate in our multiple regression model, we cannot know how this relates to previous smoking patterns (if at all). Lastly, we did not collect data on metabolic or genetic factors that are associated with the development of nicotine dependence (Audrain-McGovern et al. 2007) and may enable LITS to maintain lower levels of cigarette use.

Strategies used by LITS in this study to maintain their smoking at a low level could be useful to heavy smokers who want to achieve reduced levels of smoking. Future studies should examine how the use of smoking restraint strategies is associated with smoking trajectories over time.