Abstract
Objectives
While cessation from cigarettes is a top priority for public health, controversy surrounds the role of e-cigarettes for quitting cigarettes. This study examines the role of e-cigarettes in quit attempts and 3-month cigarette abstinence using a large, recent nationally representative US sample.
Methods
Data from the 2014/15 Tobacco Use Supplement-Current Population Survey (TUS-CPS) on cigarette and e-cigarette use and individual characteristics were supplemented with information on state tobacco control policies. We estimated frequencies and multivariate logistic equations for making a quit attempt among those who smoked 1 year earlier and for remaining abstinent at least 3 months among those making a quit attempt. These two outcomes were related to demographic characteristics, tobacco control policies and different frequency measures of e-cigarette use (ever, at least 1, 5, 20 of the last 30 days, a continuous measure of days use).
Results
Having made a quit attempt was more likely among smokers using e-cigarettes than non-users. Among those making at least one quit attempt, quit success was lower among ever users, but higher among those with at least 5 days use of e-cigarettes in the last month. Both quit attempts and quit success were linearly related to the frequency of e-cigarette use.
Conclusions
Consistent with randomized trials and those observational studies that measure frequency of e-cigarette use, both quit attempts and quit success were positively associated with increased frequency of e-cigarette use. Frequency of e-cigarette use was important in gauging the nature of these relationships.
Implications
Previous studies have obtained mixed results regarding the relationship of e-cigarette use to cigarette smoking cessation. This study provides a more precise methodology for considering the relationship of e-cigarette use to quit attempts and to quit success, and finds that quit attempts and quit success increase with the number of days use in the past month.
Introduction
The 2014 Surgeon General’s Report1 states that “the burden of death and disease from tobacco use in the United States is overwhelmingly caused by cigarettes and other combusted tobacco products; rapid elimination of their use will dramatically reduce this burden.” Accelerating smoking prevalence reduction remains one of the primary criteria for evaluating regulations under the Tobacco Control Act and the impact of policy on the public health benefit to the population as a whole.2,3
While the public health community agrees that cessation from cigarettes is a top priority, controversy surrounds the role of e-cigarettes as method for quitting cigarette use.4,5 Of five recent reviews, three reported a positive,6–8 one a negative5 and one an inconclusive9 impact of e-cigarettes on cessation. Much of the controversy surrounds observational study issues related to heterogeneity of designs, samples, and measures of exposure to the independent variables. Many studies (reviewed in detail in Glasser et al.4) are subject to: selection bias (e.g. smokers who quit by using e-cigarettes were excluded from the sample); inadequate measures of exposure (e.g. ever use in one’s lifetime, recent but infrequent use); and confounders (e.g. smokers who have repeatedly failed to quit or who are more nicotine dependent may be more likely to try e-cigarettes). The ability to control for the confounding factors that might influence e-cigarette use, i.e. whether or not those who use e-cigarettes are those who would have otherwise been likely to quit in the absence of e-cigarette use,10 can play a key role.
Imprecise measures of exposure, unmeasured confounders, having appropriate comparison groups and the outcome measures of cigarette smoking cessation (e.g. quit attempts or maintenance of cessation) are all important considerations in interpreting the mixed results from observational studies. Observational studies suggest that regular (e.g. daily use) or more intensive (e.g. use of tank or mod devices) e-cigarette use can facilitate quit attempts and cessation.11–14 Some studies with more precise measures and designs find that e-cigarette use increases quit attempts,12,15,16 while others find that e-cigarette use is associated with quit success.11,13,17
Studies of the relationship of e-cigarette use and smoking span a wide range of methodologies and samples, including uncontrolled studies, case control studies of select populations and more broad-based studies that are meant to be more representative of the population. A recent US population level study18 using a prospective follow-up from the 2010 Tobacco Use Supplement (TUS) of the Current Population Survey (CPS) found that e-cigarette use was not associated with successful cessation. However, the sample of e-cigarettes users included ever users, and consequently did not distinguish past failures, or frequency of e-cigarette use. In addition, the study was for the year 2010, and thus contained adopters of early generation e-cigarettes. Later generation models of e-cigarettes have been shown to more efficiently deliver nicotine,4,19,20 thus increasing their potential to serve as alternatives or replacements for cigarettes, including as aids in quitting cigarette use.
We applied the 2014/15 TUS-CPS to examine the relationship between frequency of e-cigarette use, quit attempts and 3-month cigarette abstinence. That survey contains information on the frequency of use (the number of days use in the last 30 days), and provides information that can be used to construct a retrospective examination of cessation and e-cigarette use in the last year. Specifically, current smokers are asked about quit attempts in the previous year and former smokers are asked when they quit, so that past year cessation behaviors can be gauged among those who were smokers 1 year ago. We considered different measures of e-cigarette use, including ever use, past 30 day use, and the number of days use.
Methods
We adopted a methodology previously applied by Levy et al.21 and we separately considered past year quit attempts and quit success. For the sample of smokers 1 year ago, we estimated frequencies and conducted multivariate logistic equations for quit attempts as a function of individual socio-demographic characteristics, tobacco control policies, cigarette use, smokeless tobacco (SLT) and e-cigarette use. For the sample of those who made at least one quit attempt in the last year, we next estimated frequencies and conducted multivariate analysis for quit success as a function of individual characteristics, policy variables, cigarette use, SLT use and e-cigarette use.
Primary Data
The primary source of data in this study was the 2014/15 TUS. The TUS is a special supplement to the CPS, and is designed to ask extensive questions on tobacco use. A probability sample employs stratified clusters of households drawn from an initial sampling frame that covers the civilian, non-institutionalized population. The 2014/15 TUS-CPS consisted of three samples collected in July 2014, January 2015, and May 2015. Each of the surveys includes about 54000 households containing over 150000 persons. The TUS-CPS asked certain tobacco use questions of both self-respondents and proxy respondents, but only self-responders, which were over 35% for each of the 3 months, were eligible to answer the more detailed tobacco-use questions. We limit the sample to ages 18 and above, self-responders, and to individuals who were smokers 1 year ago with known cigarette use.
Cigarette Smoking Cessation and Use Variables
Respondents to the TUS-CPS were first asked whether they had smoked at least 100 cigarettes in their lifetime, and then asked whether they currently smoked every day, some days, or not at all (leaving 162079 records of individuals with valid responses to these two questions). A current smoker was defined as someone who had smoked at least 100 cigarettes and who was smoking some or every day at the time of the survey. Current smokers were asked whether they were smoking cigarettes “every day, some days, or not at all around this time 12 months ago.” We included current smokers who were smoking every day or some days 12 months ago in our study samples (N = 21226). Respondents who smoked 12 days or less in the past 30 days were queried about whether they “tried to quit smoking completely during the past 12 months”. Those who smoked 12 or fewer days were considered to have made a quit attempt if they tried to quit completely. Respondents who smoked more than 12 days were queried whether they “stopped smoking for one day or longer because of trying to quit smoking during the past 12 months” and those who answered “yes” were considered to have made a quit attempt.
Former smokers were also included among those who had made a quit attempt in the past year. This group included individuals who had smoked at least 100 cigarettes and were smoking 12 months ago, but who were not smoking at the time of the survey. Former smokers were asked how long ago they quit. To capture the quitters who were more likely to remain abstinent among those who have made a quit attempt, quit success was measured by those ex-smokers who have remained abstinent for at least 3 months, since about 65% of quitters relapse within the first 3 months.22–24
Among current smokers who smoked more than 12 months (N = 21226) and former smokers who quit within 12 months (N = 3212), there was a total sample of 24438. We first omitted current smokers with unknown quit attempts (N = 147), leaving a sample of 24291 with a valid response regarding quit attempts. We also omitted former smokers who quit within the last 12 months but reported no cigarettes smoked 12 months ago (N = 626) or unknown smoking frequency 12 months ago (N = 32) from the dataset, leaving a sample of 23633 as shown in Figure 1. In the quit attempt model, the sample included those who were smokers 1 year ago (N = 23633) and the outcome is whether those smokers made a quit attempt (including those currently quit) in the last year. In the quit success model, the sample was limited to those having made a quit attempt but failed and those who quit more than 3 months (N = 10015) and the outcome is whether they succeeded (former smokers at least 3 months). Those who had quit less than 3 months were omitted from the sample, but included in the quit attempt model sample. We also considered quit success measured as smokers who quit in the last year, but not the last month.
Figure 1.
Sample Design, CPS-TUS 2014/15.
We characterized smokers by the quantity smoked per day. Those who smoked every day 12 months ago, regardless of current smoking status, were asked the average number of cigarettes they smoked per day (cpd) 12 months ago. For those who smoked some days, we weighted their response by the number of days they smoked per month 12 months ago. We classified smokers 12 months ago as very light ( < 5 cpd), light (5–14 cpd), medium (15–24 cpd), and heavy (25 or more cpd). We also included an indicator for those who consider themselves someday smokers.
E-cigarette and Smokeless Tobacco Use
Three questions were asked in the TUS-CPS about current and past use of e-cigarettes. After the respondents were given a description of e-cigarettes, they were asked, “Have you ever used e-cigarettes even one time?” Those who answer yes were asked, “Do you now use an e-cigarettes every day, some days or not at all?” Finally, respondents who used e-cigarette some days were asked, “On how many of the past 30 days did you use e-cigarettes?” We created different measures for e-cigarette use. Ever use of e-cigarettes is defined if they answer “yes” to the first question. Among those who answered “every day” and “some days” to the second question, we distinguished those who currently used e-cigarettes at least 1 day, at least 5 days, and at least 20 days in the last month. Respondents with an unknown history of e-cigarette use were treated as a separated group by creating an indicator variable as a control. In our frequency evaluations, we also considered 5 day intervals. Since SLT use may also affect cessation, corresponding measures were also developed for these users.
To examine the specific relationship of quitting behaviors to e-cigarette use, we created a continuous e-cigarette use variable, measured as the number of days of e-cigarette use in the past 30 days. In addition to the continuous measure of use, we also included a variable for at least one day use in the last 30 days to serve as the intercept term (any use). To allow for a potential non-linear relationship between days used and cessation behaviors, we examined whether the addition of squared and cubic forms of the frequency variable improved the fit of the equation. A corresponding measure was also developed for SLT users.
Socio-demographic Variables
Using socio-demographic information included in the TUS-CPS data, the sample was divided by gender, age (18–21, 22–25, 26–29, 30–34, 35–44, 45–64, 65+), and racial/ethnic groups (White, Black or African American, Asian only, and other races). A separate variable was created to distinguish Hispanics. Educational levels were classified into four groups (less than high school, high school graduate, some college but no degree, and college graduate and above). Four family income levels (less than $19999, $20000–$39999, $40000–$74999, and greater than $74999) were distinguished. We also included marital status (never married, married-spouse present, married-spouse absent, widowed/divorced/ separated), urban-rural residency, and employment status (employed vs. unemployed or not in the labor force, with the latter including those retired, those having long-term physical or mental illness, students and persons keeping house). Indicator variables were created for each of these classifications.
Policy Variables
For cigarettes, we used state prices (including generics) by year for the month of November from Orzechowski et al.25 Since monthly price data were not available, we estimated the price at the time of TUS-CPS survey (July 2014, January 2015, May 2015) by the price in November 2014 and the price in November 2015 and adjusted by the state tax change (11 states) if a change occurred between July 2014 and May 2015.26 To control for inflation, the cigarette price variable was adjusted by the consumer price index corresponding to the wave and year.27
Since smoking-free worksites have been associated with reduced smoking,28 we included an indicator for smoke-free worksites by state law. We also included e-cigarette-free worksites (only in North Dakota, New Jersey, and Utah when the survey was conducted). Since the impact is likely to depend on whether the individual works indoors, an indicator from the TUS-CPS for working indoors, but not at home is included.
Separate time indicators were included for survey months. We initially included state level indicator variables to distinguish state sentiment towards tobacco from the effect of policies, and thus better distinguish the role of policies. Based on goodness of fit and to reduce collinearity, we aggregated the state indicators into census regions: New England, Middle Atlantic, East North Central, West North Central, South Atlantic, East South Central, West South Central, Mountain, and Pacific.29
Statistical Methods
Frequencies were calculated separately for the quit attempt and quit success outcomes. A chi-square analysis of within-category differences was conducted for each of the socio-demographic factors, policies, smoking frequency, and the measures of e-cigarette/SLT use status.
Separate multivariate equations were estimated corresponding to each of the two quit measures to isolate the contribution of e-cigarette use from other variables. We conducted two-tailed tests of statistical significance, where a positive sign on the use coefficient indicates that e-cigarette use was associated with increased quit attempts and a negative sign indicates that e-cigarette use was associated with fewer quit attempts. Similarly, we conducted two-tailed tests of statistical significance to determine whether use was positively or negatively associated with quit success among those making a quit attempt.
Using the SURVEYLOGISTIC procedure of SAS (version 9.4), the quit attempt and quit success equations were estimated as weighted (using self-response weights) logistic regression models, using Fisher’s scoring method as the optimization technique. We first estimated separate equations for males and females. Based on the similarity of results, we pooled the equations, and included interactive terms for the only variable which was found to be substantively differ (widow). We also estimated separate equations by age group to focus on the ages when initiation of e-cigarettes often takes place.
Results
Mean Quit Attempts and Quit Success
Table 1 presents the means for each independent variable in terms of quit attempts among smokers and whether there was quit success among those making a quit attempt.
Table 1.
Quit Attempts among Smokers and Quit Success Among Those Making a Quit Attempt, CPS-TUS 2014/15
| Variable | Categories | Quit attempt model | Quit success model | ||||
|---|---|---|---|---|---|---|---|
| Sample size | Quit attempts§ | Chi-square | Sample size | Quit success† | Chi-square | ||
| % | (p value) | % | (p value) | ||||
| Overall | 23633 | 46.4% | 10015 | 15.9% | |||
| Gender | Male | 11767 | 44.9% | 20.73 | 4801 | 16.9% | 6.02 |
| Female | 11866 | 47.9% | <0.001 | 5214 | 15.1% | 0.014 | |
| Age | 18–21 | 555 | 57.8% | 290 | 19.0% | ||
| 22–25 | 1446 | 55.5% | 714 | 22.0% | |||
| 26–29 | 1745 | 53.8% | 838 | 18.3% | |||
| 30–34 | 2522 | 50.6% | 1162 | 18.9% | |||
| 35–44 | 4452 | 48.6% | 1972 | 15.4% | |||
| 45–64 | 9962 | 42.8% | 229.15 | 3946 | 13.2% | 56.02 | |
| 65+ | 2951 | 41.1% | <0.001 | 1093 | 17.0% | <0.001 | |
| Race | White | 19624 | 45.9% | 8198 | 17.0% | ||
| Black | 2551 | 49.0% | 1168 | 9.6% | |||
| Asian | 471 | 48.2% | 12.36 | 208 | 15.4% | 43.62 | |
| Other Races | 987 | 49.2% | 0.006 | 441 | 13.6% | <0.001 | |
| Hispanic | Hispanic | 1592 | 49.9% | 8.44 | 711 | 15.5% | 0.12 |
| Non-Hispanic | 22041 | 46.2% | 0.004 | 9304 | 16.0% | 0.725 | |
| Education | Less than 12 years | 3690 | 43.4% | 1485 | 10.0% | ||
| High school degree | 9234 | 43.3% | 3673 | 14.3% | |||
| Some college, no degree | 7688 | 49.9% | 112.22 | 3472 | 16.5% | 136.85 | |
| College degree or higher | 3021 | 50.9% | <0.001 | 1385 | 25.3% | <0.001 | |
| Family income | $0–$19999 | 6678 | 46.6% | 2882 | 11.8% | ||
| $20000–$39999 | 6625 | 45.3% | 2757 | 14.7% | |||
| $40000–$74999 | 6120 | 46.1% | 10.64 | 2539 | 17.4% | 97.12 | |
| $75000 or more | 4210 | 48.4% | 0.014 | 1837 | 22.2% | <0.001 | |
| Marital status | Never Married | 6750 | 48.6% | 2997 | 16.4% | ||
| Married Present | 8766 | 46.6% | 3714 | 18.4% | |||
| Married – Spouse Absent | 352 | 47.7% | 26.11 | 151 | 9.3% | 44.92 | |
| Widowed/Div./Sep. | 7765 | 44.4% | <0.001 | 3153 | 12.9% | <0.001 | |
| Employment | Employed | 13639 | 46.9% | 2.29 | 5796 | 17.7% | 30.78 |
| Not in labor force or unemployed | 9994 | 45.9% | 0.130 | 4219 | 13.6% | <0.001 | |
| Metropolitan status | Metropolitan | 16815 | 47.4% | 19.82 | 7249 | 16.8% | 14.70 |
| Non-metropolitan | 6818 | 44.2% | <0.001 | 2766 | 13.7% | <0.001 | |
| Indoor workers | No | 14963 | 45.7% | 9.93 | 6263 | 14.2% | 37.85 |
| Yes | 8670 | 47.8% | 0.002 | 3752 | 18.8% | <0.001 | |
| Cigarette retail price per pack | Above mean price | 8683 | 47.6% | 7.38 | 3748 | 16.5% | 1.64 |
| Below mean price | 14950 | 45.8% | 0.007 | 6267 | 15.6% | 0.200 | |
| State level worksite cigarette ban | Not highest level | 8936 | 46.3% | 0.19 | 3783 | 16.1% | 0.16 |
| Highest level | 14697 | 46.5% | 0.666 | 6232 | 15.8% | 0.688 | |
| State level worksite e-cigarette ban | Not highest level | 22760 | 46.4% | 0.01 | 9654 | 15.8% | 3.34 |
| Highest level | 873 | 46.6% | 0.909 | 361 | 19.4% | 0.068 | |
| Cigarette per day 12 months ago | 1–4 | 4426 | 57.7% | 2349 | 19.3% | ||
| 5–14 | 8075 | 48.4% | 3582 | 15.1% | |||
| 15–24 | 8361 | 40.4% | 3058 | 13.8% | |||
| 25+ | 2015 | 35.9% | 469.00 | 650 | 17.4% | 35.04 | |
| Unknown CPD | 756 | 54.2% | <0.001 | 376 | 18.4% | <0.001 | |
| Smoking frequency | Every day smokers 12 months ago | 18890 | 42.7% | 538.80 | 7292 | 14.6% | 35.47 |
| Someday smokers 12 months ago | 4743 | 61.5% | <0.001 | 2723 | 19.5% | <0.001 | |
| Smokeless use frequency in last month | Unknown use frequency | 282 | 43.6% | 108 | 16.7% | ||
| Never use | 20158 | 45.7% | 8,408 | 15.4% | |||
| 1–4 days | 111 | 41.4% | 45 | 4.4% | |||
| 5–9 days | 67 | 46.3% | 30 | 23.3% | |||
| 10–14 days | 45 | 53.3% | 23 | 4.3% | |||
| 15–19 days | 41 | 58.5% | 24 | 12.5% | |||
| 20–24 days | 22 | 54.5% | 9 | 11.1% | |||
| 25–30 days | 257 | 49.8% | 42.23 | 117 | 33.3% | 40.53 | |
| Ever, non-current use | 2650 | 51.8% | <0.001 | 1251 | 18.1% | <0.001 | |
| E-cigarette use frequency in last month | Unknown use frequency | 484 | 49.0% | 219 | 10.0% | ||
| Never use | 14469 | 39.5% | 5216 | 16.9% | |||
| 1–4 days | 698 | 51.3% | 346 | 5.2% | |||
| 5–9 days | 329 | 61.1% | 197 | 4.6% | |||
| 10–14 days | 295 | 61.4% | 166 | 6.6% | |||
| 15–19 days | 170 | 72.9% | 116 | 11.2% | |||
| 20–24 days | 120 | 72.5% | 82 | 17.1% | |||
| 25–30 days | 947 | 79.4% | 984.83 | 623 | 32.6% | 209.19 | |
| Ever, non-current use | 6121 | 54.2% | <0.001 | 3050 | 13.9% | <0.001 | |
§Quit attempts are from the sample of those who smoked 1 year ago.
†Quit success is among those who made a quit attempt and stopped smoking for at least 3 months.
Among all smokers, quit attempt rates showed significant differences for all of the individual characteristic measures, except by whether the individual was employed. The quit attempt rate also differed among smokeless tobacco users, everyday vs. someday smokers and by quantity smoked. Among the policy variables, differences were found only for the retail cigarette price. Significant differences in quit attempt rates were found for e-cigarette use at different frequencies: never use (39.5%), ever but not current use (54.2%), and for 1–4 days (51.3%), 5–9 days (61.1%), 10–14 days (61.4%), 15–19 days (72.9%), 20–24 days (72.5%), and at least 25 days in the last month (79.4%).
Among all smokers 12 months ago who made at least one quit attempt during the past 12 months, quit success rates showed significant differences among individual characteristics, except whether Hispanic. Quit success rates differed by smoking cigarette frequency, cigarette quantity smoked and whether used smokeless tobacco. Significant differences were observed in quit success between e-cigarette use at different frequencies: never use (16.9%), ever but not current use (13.9 %), and for 1–4 days (5.2 %), 5–9 days (4.6 %), 10–14 days (6.6 %), 15–19 days (11.2 %), 20–24 days (17.1 %), and at least 25 days in the last month (32.6%).
Logistic Regression
Tables 2 and 3 reports the results for the coefficients of the cigarette, SLT and e-cigarette use variables in the quit attempt and quit success logistic regression equations. The complete results, including the coefficients of the socio-demographic, policy, region and time indicator variables, are contained in Supplement 1. The different measures of e-cigarette and smokeless tobacco use are distinguished by column.
Table 2.
Logistic Regression Analysis of Having Made a Quit Attempt among a Sample of Smokers 1 Year Ago (N = 23,633), CPS-TUS 2014/15
| Variable (reference variable in parentheses) | Categories | Ever use | Current use | 5+ days in last 30 days | 20+ days in last 30 days | Days of use in last 30 days |
|---|---|---|---|---|---|---|
| AOR (LL, UL) sig.§ | AOR (LL, UL) sig.§ | AOR (LL, UL) sig.§ | AOR (LL, UL) sig.§ | AOR (LL, UL) sig.§ | ||
| Cigarettes per day 12 months ago (1–4) | 5–14 | 0.99 (0.87, 1.12) | 1.04 (0.92, 1.17) | 1.04 (0.92, 1.17) | 1.05 (0.93, 1.18) | 1.04 (0.92, 1.17) |
| 15–24 | 0.74 (0.65, 0.85)*** | 0.79 (0.70, 0.91)*** | 0.80 (0.70, 0.92)** | 0.81 (0.71, 0.93)** | 0.80 (0.70, 0.91)*** | |
| 25+ | 0.63 (0.54, 0.75)*** | 0.70 (0.59, 0.83)*** | 0.70 (0.60, 0.83)*** | 0.72 (0.61, 0.85)*** | 0.70 (0.59, 0.83)*** | |
| Unknown CPD | 0.96 (0.79, 1.17) | 0.96 (0.79, 1.16) | 0.95 (0.78, 1.16) | 0.95 (0.78, 1.15) | 0.95 (0.78, 1.16) | |
| Smoking frequency (every day smoker) | Someday smoker | 1.71 (1.52, 1.92)*** | 1.72 (1.52, 1.93)*** | 1.72 (1.53, 1.94)*** | 1.74 (1.55, 1.96)*** | 1.73 (1.54, 1.95)*** |
| Smokeless use† | Yes | 1.03 (0.92, 1.14) | 0.87 (0.69, 1.09) | 0.99 (0.76, 1.28) | 0.95 (0.69, 1.30) | 0.72 (0.48, 1.07) |
| Unknown | 0.59 (0.40, 0.85)** | 0.58 (0.40, 0.84)** | 0.58 (0.40, 0.84)** | 0.59 (0.40, 0.86)** | 0.58 (0.40, 0.85)** | |
| E-cigarette use† | Yes | 2.31 (2.15, 2.48)*** | 2.60 (2.32, 2.90)*** | 3.39 (2.97, 3.87)*** | 4.90 (4.09, 5.85)*** | 1.25 (1.05, 1.49)* |
| Unknown | 2.03 (1.55, 2.66)*** | 1.65 (1.26, 2.15)*** | 1.63 (1.25, 2.13)*** | 1.58 (1.21, 2.07)*** | 1.64 (1.25, 2.15)*** | |
| Smokeless use by number of days in last 30 days | 1.01 (0.99, 1.03) | |||||
| E-cigarette use by number of days in last 30 days | 1.05 (1.04, 1.06)*** | |||||
| Model Fit | Wald Statistic | 23.18*** | 17.68*** | 18.87*** | 18.55*** | 18.83*** |
§AOR: adjusted odds ratio. (LL, UL): lower and upper limit for 95% confidence interval. Significance codes for null hypothesis (AOR = 0): 0. “***” 0.001; “**” 0.01; “*” 0.05; “.” 0.1.
†Smokeless and e-cigarette use were grouped by “Yes”, “No”, and “Unknown” in the first four models (Ever use, Current use, 5 or more, and 20 or more days in last 30 days) and were grouped by “Current use”, “Non-current use”, “Unknown”, with a continuous variable for specific number of days of use in last 30 days in the last model.
Table 3.
Logistic Regression Analysis of Quit Success (Abstinent at Least 3 Months) among a Sample of Those Making a Quit Attempt in the Last Year (N = 10,015), CPS-TUS 2014/15
| Variable (reference variable in parentheses) | Categories | Ever use | Current use | 5+ days in last 30 days | 20+ days in last 30 days | Days of use in last 30 days |
|---|---|---|---|---|---|---|
| AOR (LL, UL) sig.§ | AOR (LL, UL) sig.§ | AOR (LL, UL) sig.§ | AOR (LL, UL) sig.§ | AOR (LL, UL) sig.§ | ||
| Cigarettes per day 12 months ago (1–4) | 5–14 | 0.91 (0.72, 1.13) | 0.88 (0.70, 1.10) | 0.87 (0.70, 1.09) | 0.85 (0.68, 1.07) | 0.86 (0.69, 1.08) |
| 15–24 | 0.90 (0.70, 1.16) | 0.85 (0.66, 1.10) | 0.84 (0.65, 1.08) | 0.82 (0.63, 1.05) | 0.84 (0.65, 1.09) | |
| 25+ | 1.24 (0.88, 1.73) | 1.13 (0.81, 1.58) | 1.10 (0.78, 1.53) | 1.04 (0.74, 1.45) | 1.09 (0.78, 1.54) | |
| Unknown CPD | 0.98 (0.69, 1.38) | 0.97 (0.68, 1.38) | 0.97 (0.68, 1.38) | 0.97 (0.68, 1.38) | 0.95 (0.67, 1.35) | |
| Smoking frequency (every day smokers 12 months ago) | Someday smokers 12 months ago | 1.23 (1.00, 1.52). | 1.22 (0.99, 1.51). | 1.21 (0.98, 1.50). | 1.22 (0.99, 1.51). | 1.24 (1.00, 1.53)* |
| Smokeless use† | Yes | 1.06 (0.87, 1.29) | 1.26 (0.85, 1.87) | 1.50 (0.99, 2.28). | 2.00 (1.21, 3.31)** | 0.51 (0.19, 1.38) |
| Unknown | 1.78 (0.95, 3.33). | 1.76 (0.94, 3.29). | 1.75 (0.93, 3.31). | 1.77 (0.92, 3.43). | 1.87 (1.00, 3.53). | |
| E-cigarette use† | Yes | 0.80 (0.69, 0.92)** | 1.22 (1.02, 1.46)* | 1.59 (1.31, 1.92)*** | 2.81 (2.26, 3.49)*** | 0.17 (0.10, 0.28)*** |
| Unknown | 0.47 (0.29, 0.74)** | 0.54 (0.34, 0.85)** | 0.56 (0.35, 0.88)* | 0.57 (0.36, 0.91)* | 0.52 (0.33, 0.82)** | |
| Smokeless use by number of days in last 30 days | 1.05 (1.01, 1.09)* | |||||
| E-cigarette use by number of days in last 30 days | 1.10 (1.08, 1.12)*** | |||||
| Model Fit | Wald Statistic | 6.16*** | 6.01*** | 6.41*** | 7.63*** | 8.41*** |
§AOR: adjusted odds ratio. (LL, UL): lower and upper limit for 95% confidence interval. Significance codes for null hypothesis (AOR = 0): 0. “***” 0.001; “**” 0.01; “*” 0.05; “.” 0.1.
†Smokeless and e-cigarette use were grouped by “Yes”, “No”, and “Unknown” in the first four models (Ever use, Current use, 5 or more, and 20 or more days in last 30 days) and were grouped by “Current use”, “Non-current use”, “Unknown”, with a continuous variable for specific number of days of use in last 30 days in the last model.
Quit Attempts by Smokers
Among smokers 12 months ago, having made at least one quit attempt was generally more likely among those at lower ages (age 18–44), African Americans, those with higher education (some college but no degree, college degree or higher), and someday smokers. A quit attempt was less likely among those with income of $20000–$39999, the employed, those never married, non-metropolitan residents, and among those heavy smokers (smoking 15–24 and at least 25 cigarettes per day). Compared to those never married, males who are widowed, divorced or separated were less likely to make a quit attempts than females.
While all measures of e-cigarette use showed a significant relationship (p < 0.001) with having made a quit attempt, the adjusted odds ratios (AOR) increased continuously with e-cigarette frequency from 2.3 for ever use to 4.9 for 20+ days use in the last month. For the continuous use variable, the linear use variable was significant, but the quadratic and cubic terms were not significant and reduced overall equation fit in terms of the Wald statistic. With only the linear use variable, the AOR had a value of 1.05, indicating the odds of a quit attempt increased by 5% with each additional day of use. None of the SLT use variables, continuous or otherwise, were significant. Similar results (not shown) in terms of the sign of the coefficient and level of significance for the e-cigarette and SLT use variables were found for males and females when we estimated separate equations by gender, and for those ages 18–34 when distinguished from other ages.
Quit Success among Smokers Making a Quit Attempt
For smokers having made a quit attempt, including those who failed to quit and those who quit smoking for at least 3 months, quit success was higher among those at age 18–34 and age 65 and above compared with those ages 35–64, those with more education (some college but no degree, college degree or higher) and higher income ($75000 or more), and someday smokers compared to every day smokers. Quit success was less likely among Blacks, those never married, those married but spouse absent, those widowed, divorced or separated, and non-metropolitan residents. Among the policy variables, no significant associations were found.
A negative relationship was obtained between quit success and e-cigarette ever use, but a positive relationship was observed with current e-cigarette use measures. Moreover, the AOR of quit success was 59% higher with 5+ days use and 181% higher with 20+ days use. For the continuous use variable, the linear use variable quit was significant, but the quadratic and cubic terms were not significant and worsened overall fit. In linear form, the AOR had a value of 1.10, indicating the odds of quit success increase by 10% with each additional day of use. Quit success was found to be positively related to SLT use for 20+ days use (AOR = 2.00) and continuous use (AOR = 1.05). Similar results (not reported) in terms of the sign of the coefficient and level of significance of the e-cigarette and SLT use were obtained for males and females and for those ages 18–34 as obtained for the whole sample.
Discussion
Unlike an earlier study using the 2010 TUS-CPS18 and some earlier observational studies that employed heterogeneous and imprecise measures of e-cigarette frequency of use,4,5 we found that more precise measures of use yielded a common and consistent set of results. Results revealed that greater frequency of e-cigarette use beyond ever use and especially with 20 or more days of use in the past month was strongly associated with both having made a quit attempt and a greater likelihood of 3 months or more of cigarette smoking cessation. In particular, we found that quit success was negatively related to ever use, but positively related to 5 or more days use in the last month. The consistency of positive associations with quit attempts or cessation success suggests that more frequent e-cigarette use may be effective as a smoking cessation aid.
Amato et al.30 reported that e-cigarette use of 6 or more days in the previous month was more likely to be related to use for the purposes of quitting cigarette smoking. Also consistent with our results, Biener and Hargrave11 reported that daily use of e-cigarettes for a month or more was associated with 6 times greater likelihood of smoking cessation measured 2 years later, whereas ever use or use on a few occasions was either unrelated to cessation or negatively associated with cessation. In recent national surveys conducted during approximately the same time periods as our TUS-CPS results, recent quitters are nearly four times more likely to be exclusive daily users of e-cigarettes than current smokers (13.0% vs. 3.5%, respectively).31 A similar relationship is found with almost two times as many daily e-cigarette users among smokers who had quit smoking 2 to 3 years ago.31,32 Our results are also consistent with the few randomized trials6–8 and the better controlled observational studies12–14 that suggest more intensive patterns of e-cigarette use can encourage quit attempts or quit success. Since our sample contained those who currently or previously smoked, these results suggest the importance of frequency in measures of dual (both e-cigarette and cigarette) use. It is noteworthy that previous studies with negative associations of e-cigarette use to cessation generally did not measure the frequency of exposure and some only included ever e-cigarette use in one’s lifetime.5
An important part of our study was to distinguish the effect of e-cigarette use on quit attempts and quit success by separately considering the smokers who might attempt to quit and those that have actually made a quit attempt. While some early studies found a relationship with either quit attempts or quit success,11–13,15–17 the large TUS-CPS sample enabled us to distinguish effects on both quit attempts and quit success. Our results indicate that e-cigarettes played a greater role (i.e. had a higher odd ratio) in getting smokers to make a quit attempt than in quit success. We obtained similar results for those ages 18–34 suggesting that e-cigarette use may be especially beneficial at ages when quit success tends to be low.
We also examined the relationship between e-cigarette use and quit success using a 1-month rather than 3-month cut-off measure, and obtained consistent results. Different questions were asked in the CPS-TUS survey about quit attempts for those who smoked 12 or less days in the last month compared to those who smoker more than 12 days. Those who had smoked more than 12 days in the past 30 days and had not stopped for one day or longer to quit smoking were also asked if they “made a serious attempt to stop smoking because of trying to quit - even if [s/he] stopped for less than one day.” When we included this group as having made a quit attempt, their inclusion did not substantively affect the results.
An important part of the study was to control for the factors that are likely to affect quit attempts and quit success, including not only individual characteristics, but also smoking characteristics. The relationship of days used to quitting behaviors changed, e.g. for the quit success variable, when we controlled for other variables in the logistic regression equations (Table 3) as compared to the simple frequencies (Table 1). The importance of SLT use diminished (p > 0.1) in the quit attempt model compared to simple frequencies and the importance of cigarette quantity diminished in the quit success model compared to the simple frequencies. While the results presented above included measures of someday versus every day and quantity smoked, we also estimated equations which included indicators of the time after awakening until smoking and of menthol cigarette use. The results for e-cigarette use changed little with this these variables included. We also considered an indicator variable based on a survey question to smokers regarding whether they tried to quit by switching to e-cigarette. That variable was often negative when included alone and was not significant when added to the multivariate logistic equations reported above.
The results were also robust to different measures of the policy variables, including cigarettes taxes rather than price, individual worksite restrictions rather than state worksite laws, and to the addition of a variable measuring state per capita tobacco control spending. Policy variables were generally not significant. However, the effect of e-cigarette smoke-free air laws merits further consideration, since the laws were adopted in only three states soon before the time of the survey.
We also considered smokeless tobacco use. We did not find a relationship to having made a quit attempt, but we did find that high rates of use were related to quit success.
Despite measuring frequency of e-cigarette use, controlling for many factors that may influence cessation, and having a large, representative sample, this study has limitations. Since our study is cross-sectional and used retrospective data over the past year, it is subject to the following limitations: recall bias; an inability to discern the temporality of precisely when the specific pattern of e-cigarette use and the quit attempt or quit episode of greater than 3 months occurred, and therefore whether use was most proximal to and preceded the quit attempt /3 month abstinent outcome; and omission of plausible confounders, cannot be ruled out.4 Among plausible omitted confounders are the type of e-cigarette used (tank vs. cigalike), attempts to quit in previous years (with or without pharmacotherapy, and with or without e-cigarettes), or other factors which may reflect motivation to quit. Indeed, it is difficult to distinguish whether the increased probability of quitting is due to the frequent use of the product or that frequent users are more committed to quitting smoking, i.e. the motivations of the frequent users. In addition, our results may be understated to the extent that some smokers may have used e-cigarettes in the last year and quit both e-cigarettes and cigarettes, but may be overstated to the extent that those who quit smoking using e-cigarettes may be more prone to later relapse. Finally, our use measure focused on the number of days use in the last 30 days, and does not incorporate use over more prolonged periods of time than the previous month nor the intensity of daily use, i.e. the number of e-cigarette use occasions per day or the type of device. Future study should consider these other characteristics of use.
Despite this study’s limitations, the more precise measures of frequency of e-cigarette use exposure and cessation outcomes appear to be a common thread across this and the other observational and RCT studies that show a positive association of greater frequency of e-cigarette use and cessation. Our results also show a strong positive association between greater frequency of e-cigarette use and both quit attempts and quit success. Nevertheless, further analysis is warranted. High quality randomized control trials with appropriate control groups and “real world” observational studies4 are needed with sufficiently rigorous measures of exposure, plausible confounders and representative samples. In addition, it will be important to monitor this relationship for future cohorts of smokers that have different patterns of past use and exposures to e-cigarette availability, and to consider the type of product as new e-cigarette or heat-no-burn cigarettes come onto the market.
Supplementary Material
Supplementary data are available at Nicotine and Tobacco Research online.
Funding
Funding was received by Drs. Levy and Abrams from the National Institute on Drug Abuse under grant R01DA036497. Dr. Levy also received funding from the National Cancer Institute under grant P01-CA200512.
Declaration of Interests
None declared.
Supplementary Material
Acknowledgments
We would like to thank Ron Borland of The Cancer Council Victoria, K. Michael Cummings of the Medical University of South Carolina, Maciej Goniewicz of the University of Buffalo, and Eric Lindblom, Georgetown University Law School, for their comments on an earlier draft.
References
- 1. U.S. Department of Health and Human Services. The Health Consequences of Smoking—50 Years of Progress: A Report of the Surgeon General. Atlanta Georgia: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, Office on Smoking and Health; 2014. [Google Scholar]
- 2. Villanti AC, Vargyas EJ, Niaura RS, Beck SE, Pearson JL, Abrams DB. Food and Drug Administration regulation of tobacco: integrating science, law, policy, and advocacy. Am J Public Health. 2011;101(7):1160–1162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Zeller M, Hatsukami D; Strategic Dialogue on Tobacco Harm Reduction Group The Strategic Dialogue on Tobacco Harm Reduction: a vision and blueprint for action in the US. Tob Control. 2009;18(4):324–332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Glasser AM, Collins L, Pearson JL et al. . Overview of Electronic Nicotine Delivery Systems: A Systematic Review. Am J Prev Med. 2017;52(2):e33–e66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Kalkhoran S, Glantz SA. E-cigarettes and smoking cessation in real-world and clinical settings: a systematic review and meta-analysis. Lancet Respir Med. 2016;4(2):116–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Hartmann-Boyce J, McRobbie H, Bullen C, Begh R, Stead LF, Hajek P. Electronic cigarettes for smoking cessation. Cochrane Database Syst Rev. 2016;9:CD010216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. McRobbie H, Bullen C, Hartmann-Boyce J et al. . Electronic cigarettes for smoking cessation and reduction. Cochrane Database Syst Rev. 2014;12:Cd010216. [DOI] [PubMed] [Google Scholar]
- 8. Rahman MA, Hann N, Wilson A, Mnatzaganian G, Worrall-Carter L. E-cigarettes and smoking cessation: evidence from a systematic review and meta-analysis. PLoS One. 2015;10(3):e0122544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Malas M, van der Tempel J, Schwartz R et al. . Electronic Cigarettes for Smoking Cessation: A Systematic Review. Nicotine Tob Res. 2016;18(10):1926–1936. [DOI] [PubMed] [Google Scholar]
- 10. Levy DT, Cummings KM, Villanti AC et al. . A framework for evaluating the public health impact of e-cigarettes and other vaporized nicotine products. Addiction. 2017; 112(1):8–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Biener L, Hargraves JL. A longitudinal study of electronic cigarette use among a population-based sample of adult smokers: association with smoking cessation and motivation to quit. Nicotine Tob Res. 2015;17(2):127–133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Brose LS, Hitchman SC, Brown J, West R, McNeill A. Is the use of electronic cigarettes while smoking associated with smoking cessation attempts, cessation and reduced cigarette consumption? A survey with a 1-year follow-up. Addiction. 2015;110(7):1160–1168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Brown J, Beard E, Kotz D et al. . Real-world effectiveness of e-cigarettes when used to aid smoking cessation: a cross-sectional population study. Addiction. 2014;109(9):1531–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hitchman SC, Brose LS, Brown J, Robson D, McNeill A. Associations Between E-Cigarette Type, Frequency of Use, and Quitting Smoking: Findings From a Longitudinal Online Panel Survey in Great Britain. Nicotine Tob Res. 2015;17(10):1187–1194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Nayak P, Pechacek TF, Weaver SR, Eriksen MP. Electronic nicotine delivery system dual use and intention to quit smoking: Will the socioeconomic gap in smoking get greater?Addict Behav. 2016;61:112–116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Filippidis FT, Laverty AA, Vardavas CI. Experimentation with e-cigarettes as a smoking cessation aid: a cross-sectional study in 28 European Union member states. BMJ Open. 2016;6(10):e012084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Beard E, West R, Michie S, Brown J. Association between electronic cigarette use and changes in quit attempts, success of quit attempts, use of smoking cessation pharmacotherapy, and use of stop smoking services in England: time series analysis of population trends. BMJ. 2016;354:i4645. [DOI] [PubMed] [Google Scholar]
- 18. Shi Y, Pierce JP, White M et al. . E-cigarette use and smoking reduction or cessation in the 2010/2011 TUS-CPS longitudinal cohort. BMC Public Health. 2016;16(1):1105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Farsalinos KE, Spyrou A, Tsimopoulou K, Stefopoulos C, Romagna G, Voudris V. Nicotine absorption from electronic cigarette use: comparison between first and new-generation devices. Sci Rep. 2014;4:4133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Wagener TL, Floyd EL, Stepanov I et al. . Have combustible cigarettes met their match? The nicotine delivery profiles and harmful constituent exposures of second-generation and third-generation electronic cigarette users. Tob Control. 2016;26(e1):e23–e28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Levy DT, Romano E, Mumford E. The relationship of smoking cessation to sociodemographic characteristics, smoking intensity, and tobacco control policies. Nicotine Tob Res. 2005;7(3):387–396. [DOI] [PubMed] [Google Scholar]
- 22. Gilpin EA, Pierce JP, Farkas AJ. Duration of smoking abstinence and success in quitting. J Natl Cancer Inst. 1997;89(8):572–576. [DOI] [PubMed] [Google Scholar]
- 23. Hughes JR, Keely J, Naud S. Shape of the relapse curve and long-term abstinence among untreated smokers. Addiction. 2004;99(1):29–38. [DOI] [PubMed] [Google Scholar]
- 24. U.S. Department of Health and Human Services. The Health Benefits of Smoking Cessation: a report of the Surgeon General. Atlanta Georgia: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, Office on Smoking and Health; 1990. [Google Scholar]
- 25. Orzechowski and Walker. The tax burden on tobacco, historical compilation. Volume 43. Arlington, VA; 2016.[AU: Please provide publisher details for reference 25 [Google Scholar]
- 26. Tobacconomics website. State Taxes. Chicago, Tobacconomics; 2016. www.tobacconomics.org. 2016. Accessed March 8, 2017. [Google Scholar]
- 27. US Department of Labor: Bureau of Labor Statistics. Consumer Price Index 2015. http://www.bls.gov/cpi/#tables. Accessed January 28, 2015.
- 28. Callinan JE, Clarke A, Doherty K et al. . Legislative smoking bans for reducing secondhand smoke exposure, smoking prevalence and tobacco consumption. Cochrane Database Syst Rev. 2010;(4):CD005992. [DOI] [PubMed] [Google Scholar]
- 29. U.S. Bureau of the Census. Census Regions and Divisions of the United States. Washington, DC, U.S. Bureau of the Census; 2016. http://www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf. Accessed July 2, 2016. [Google Scholar]
- 30. Amato MS, Boyle RG, Levy D. How to define e-cigarette prevalence? Finding clues in the use frequency distribution. Tob Control. 2016;25(e1):e24–e29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Delnevo CD, Giovenco DP, Steinberg MB et al. . Patterns of Electronic Cigarette Use Among Adults in the United States. Nicotine Tob Res. 2016;18(5):715–719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Coleman BN, Rostron B, Johnson SE, Ambrose BK et al. . Electronic cigarette use among US adults in the Population Assessment of Tobacco and Health (PATH) Study, 2013–2014 [published online ahead of print June 17, 2017]. Tob Control. 2017;pii:tobaccocontrol-2016–053462. doi:10.1136/tobaccocontrol-2016–053462. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.