Abstract
Objectives:
As of October 2015, evidence needed to make a recommendation about the use of electronic nicotine delivery systems (ENDS) for smoking cessation was limited. We used the 2014 Arkansas Behavioral Risk Factor Surveillance System with additional state-specific questions to determine the prevalence of ENDS use, the impact of ENDS use on smoking cessation, and beliefs about ENDS use in Arkansas. Our objectives were to determine if (1) ENDS use was associated with lower odds of quitting smoking, (2) ENDS users believed that ENDS use was not harmful to their health, and (3) ENDS users believed that switching to ENDS reduced their tobacco-related health risks.
Methods:
We conducted a cross-sectional study of 4465 respondents to the Arkansas Behavioral Risk Factor Surveillance System and used weighted analyses to account for the complex survey design. We used a subset of records formed by (1) formers smokers who quitted smoking in the last 5 years and (2) current smokers to assess the odds of quitting.
Results:
In 2014, 6.1% (95% confidence interval [CI], 5.0%-7.4%) of Arkansas adults were currently using ENDS. Of the 1083 participants who were current smokers or had quit smoking within the past 5 years, 515 (54.1%) had used ENDS. Of the 515 ENDS users, 404 (80.3%) had continued smoking. ENDS use was significantly associated with reduced odds of quitting smoking (weighted odds ratio = 0.53; 95% CI, 0.34-0.83). Although 2437 of 3808 participants (62.5%) believed that it was harmful for nonsmokers to start using ENDS and 1793 of 3658 participants (47.0%) believed that switching to ENDS did not reduce tobacco-related health risks, only 80 of 165 (41.3%) and 50 of 168 (33.9%) ENDS users shared these same respective beliefs.
Conclusions:
Most smokers who indicated smoking in the past 5 years and who tried ENDS did not stop smoking. ENDS use was inversely associated with smoking cessation. Tobacco cessation programs should tell cigarette smokers that ENDS use may not help them quit smoking.
Keywords: Arkansas, beliefs, electronic cigarettes, prevalence, smoking cessation
Electronic nicotine delivery systems (ENDS) were introduced in the United States in 2007.1 ENDS refers to disposable and refillable electronic cigarettes, known as e-cigarettes, and newer devices, such as cartomizers and atomizers. By 2015, an estimated 24.1% of high school students in the United States were current ENDS users, making ENDS the most commonly used tobacco product among teenagers, whereas current cigarette smoking was at 10.8% among this population.2 The prevalence of ENDS use among US adults was estimated at 4.2% during 2012 and 2013 and 3.7% in 2014.3,4
Other than reports from Kansas (2014)5 and Hawaii (2013),6 most data on ENDS use in the United States are available only at the national level. For many years, data on cigarette smoking have been collected by all US states through the Behavioral Risk Factor Surveillance System (BRFSS), a national public health surveillance system for monitoring public health status and risk behaviors for chronic disease and injury.7 The national BRFSS does not include questions about ENDS use. However, in 2014, Arkansas included questions about ENDS use as part of its BRFSS.
Arkansas is a predominantly rural state that borders the southern and midwestern regions; as such, it intersects the 2 US regions with the highest prevalences of smoking.8,9 The prevalence of current cigarette smoking in Arkansas in 2014 (24.7%) was substantially higher than the US median (18.1%).10 Because of its high prevalence of smoking, Arkansas has a substantial stake in tobacco-related public health education and efforts to reduce the use of tobacco products.
As state and national public health programs have publicized the risks of cigarette smoking, the tobacco industry has marketed ENDS as an alternative option that aids in smoking cessation11 and is purportedly “safer” than cigarette smoking.12 However, a meta-analysis of the value of ENDS in smoking cessation13 and a systematic review on the safety of ENDS use14 suggest a lack of scientific support for these claims. Furthermore, data are lacking on what the general public and ENDS users believe about these 2 issues.
We used the new ENDS-related data from the 2014 Arkansas BRFSS to determine the prevalence of ENDS use and its impact on smoking cessation in Arkansas. At the time that the Arkansas BRFSS ENDS data were made available to us, evidence on the benefits of ENDS use as a cessation device was limited and conflicting.15 We also used standard Arkansas BRFSS smoking data from 2011 to 2014 to determine changes in cigarette smoking prevalence during the time that ENDS use was increasing. Our primary objective was to determine whether ENDS use would decrease the likelihood of quitting smoking. Our secondary objective was to establish if ENDS users believed that (1) switching to ENDS reduces their tobacco-related health risks and (2) ENDS are not harmful to their health.
Methods
Data Source
In 2015, we analyzed the data available on ENDS use from the 2014 Arkansas BRFSS. The BRFSS is a standardized, random-digital-dial health survey conducted by the 50 US states, Washington, DC, and Puerto Rico, the Virgin Islands, and Guam under the guidance of the Centers for Disease Control and Prevention. Data are collected from a representative sample of adults in each state, and the sampling design provides state estimates. The BRFSS enables states to track key health indicators and risk factors for chronic disease and injury in adults. A description of the methods used for data collection is published elsewhere.7 The BRFSS contains core questions that are asked annually and other questions that are asked regularly but less frequently. States can also add questions to the survey. The Arkansas Center for Health Statistics has been conducting a monthly BRFSS survey continuously since 1993 and currently completes more than 4000 landline telephone and 1200 cellular telephone interviews annually.16 This study was exempt from human subjects review because we used de-identified secondary data.
Measures
Participants were asked, “Have you smoked at least 100 cigarettes in your entire life?” Those responding yes were considered “ever cigarette smokers” and were subsequently asked, “Do you now smoke cigarettes every day, some days, or not at all?” to determine the numbers of current smokers (ie, those who smoked every day or some days) and former smokers. Current smokers were asked, “During the past 12 months, have you stopped smoking for 1 day or longer because you were trying to quit smoking?” to identify the number of smokers who had tried to quit cigarette smoking. Former smokers were asked, “How long has it been since you last smoked a cigarette, even 1 or 2 puffs?” to quantify the time since quitting smoking. Finally, current smokers were asked if they had tried to switch to other tobacco products, including ENDS.
Questions about ENDS use were prefaced with the following statement: “The next set of questions that I am going to ask concern electronic cigarettes. Electronic cigarettes, or e-cigarettes as they are often called, are battery-operated devices that simulate smoking a cigarette. The heated vapor produced by an e-cigarette often contains nicotine.” Participants were then asked, “Have you ever used an electronic cigarette in your entire life?” If they answered yes, then they were asked, “Have you used an electronic cigarette in the past 30 days?” and “Have you used an electronic cigarette daily?” Those who reported using ENDS in the past 30 days were also asked if they had used flavored ENDS at any time. We defined current ENDS use as having used ENDS in the past 30 days. Everyday ENDS users were those who reported using ENDS systems daily.
Arkansas BRFSS also asked participants about their beliefs on the potential health benefits of ENDS use for nonsmokers and smokers. Participants were asked about their degree of agreement or disagreement with 2 statements: “If a nonsmoker begins to use electronic cigarettes, it is harmful to their health” and “If a current cigarette smoker switches to electronic cigarettes, it does not reduce their tobacco-related health risks.” Their answers were recorded on a 5-item Likert scale, where 1 = strongly agree and 5 = strongly disagree.
To determine whether the introduction of ENDS was associated with a change in the prevalence of current and former cigarette smoking, we obtained prevalence data on cigarette smoking from the Arkansas BRFSS for 1996 through 2014. In 2011, the BRFSS implemented a new survey design in which participants were contacted by landline only, cellular telephone only, or combined landline and cellular telephones. To ensure comparability of the data, we used only the BRFSS cigarette smoking prevalence data obtained after that change (ie, 2011-2014).
Statistical Methods
We included all 5258 participants of the 2014 BRFSS survey in our study because deleting records with missing or irrelevant data for some of the analysis might have resulted in bias. In addition, we comprised a subset of interest to assess the odds of quitting smoking by including 1083 participants who were current smokers or had quit smoking within the past 5 years. For analysis of the belief that ENDS use is harmful to nonsmokers’ health, we used 3808 of 5258 records (72.4%) that had complete data. Missing data on this or other key variables (eg, race/ethnicity, education, smoking status) were found in 797 (15.2%) respondents; in addition, 623 (11.8%) responded “do not know,” and 30 (0.6%) refused to answer. To analyze the belief that switching to ENDS does not reduce tobacco-related health risks for current smokers, 3658 (69.6%) respondents had complete data. Missing data on this or other key variables were found in 803 (15.3%) records; 743 (14.1%) responded “do not know” and 54 (1.0%) did not answer the question.
We estimated proportions and prevalences using the weighted totals, and we calculated 95% confidence intervals (CIs) using estimates of the variance obtained by Taylor series linearization. We conducted simple stratified analyses, and we calculated weighted prevalence ratios (PRs) and weighted odds ratios (ORs) with 95% CIs. We tested for homogeneity of weighted PRs using the Breslow-Day test. We modeled current ENDS use by using multivariate logistic regression analysis, which controlled for age (18-24, 25-34, 45-54, 55-64, ≥65), sex, race/ethnicity (non-Hispanic white, non-Hispanic black/Hispanic, non-Hispanic other, and non-Hispanic multiracial), and education (<high school, high school, some college/technical school, and ≥college or technical school graduate).
We conducted separate multivariate dichotomous logistic regression analyses restricted to the subset of former smokers who stopped smoking during the past 5 years and current smokers. In these analyses, the exposure variable was ENDS use (current and ever), and the outcome variable was quitting smoking (yes or no), controlling for age, sex, race/ethnicity, and education. We also used multivariate multinomial logistic regression analysis to determine any association between ENDS use and time since quitting smoking, given that the outcome variable had various intervals of time since quitting.
We conducted a sensitivity analysis by fitting the models to data that included the records of those who did not respond to the questions about ENDS use, as if they had either ever or never used ENDS. For the 2 statements on beliefs about ENDS, we regressed the 5 possible responses (levels of agreement or disagreement) as 2 ordinal variables in proportional odds models, where cigarette smoking, ENDS use, age, sex, race/ethnicity, and education were entered as predictors. We conducted this regression analysis using the technique described by McCullagh,17 and we applied the Satterthwaite-adjusted F test (the most conservative of the tests available in SUDAAN) to evaluate the significance of each predictor. The models predicted the weighted prevalences in our results. We used the algorithms available in SAS-callable SUDAAN release 11.0 for all statistical analyses.18 We considered P ≤ .05 to be significant.
We used smoking data from the Arkansas BRFSS to compare prevalence of cigarette smoking over time. We tested for linear trends in cigarette smoking prevalence from 2011 to 2014 using a logistic model that contained a linear time variable plus age. We also used the National Cancer Institute Joinpoint Regression Program19 to assess cigarette smoking trends, testing the hypothesis of no change (β = 0) in the prevalence of current smoking “every day” and “some days” and former smoking during the 4-year period.
We examined all 5258 records from the 2014 Arkansas BRFSS. We primarily used only the records of the 4465 participants who provided complete responses for our variables of interest. A total of 1083 of the 4465 participants were current smokers or had quit smoking within the past 5 years.
Results
Prevalence of ENDS Use
Of the 4465 study participants with complete responses for the variables of interest, 592 (weighted prevalence = 20.6%; 95% CI, 18.6%-22.8%) reported trying ENDS in their lifetime, 179 (weighted prevalence = 6.1%; 95% CI, 5.0%-7.4%) reported current ENDS use (in the past 30 days), and 80 (weighted prevalence = 2.5%; 95% CI, 1.9%-3.5%) reported daily ENDS use. Of the 179 current ENDS users, 128 (weighted prevalence = 74.4%) were current smokers, 46 (weighted prevalence = 23.8%) were former smokers, and 5 (weighted prevalence = 1.8%) had never smoked (Table 1).
Table 1.
Weighted prevalences and prevalence ratios of current ENDS use among adults, by demographic characteristics, education, and cigarette smoking status, Arkansas Behavioral Risk Factor Surveillance System, 2014
| Characteristic | Current ENDS Users, No. (%) | Study Participants, No. | Weighteda Prevalence of Current ENDS Use, % (95% CI) | Multivariate Weighteda Prevalence Ratiob (95% CI) |
|---|---|---|---|---|
| Total | 179 (100.0) | 4465 | 6.1 (5.0-7.4) | |
| Age, y | ||||
| 18-24 | 12 (12.5) | 137 | 6.8 (3.7-12.5) | 3.7 (1.7-8.2) |
| 25-34 | 19 (21.9) | 295 | 8.0 (4.9-12.7) | 3.6 (1.8-7.0) |
| 35-44 | 33 (25.2) | 374 | 9.3 (6.2-13.7) | 4.2 (2.2-8.0) |
| 45-54 | 39 (20.5) | 613 | 7.3 (4.9-10.8) | 3.2 (1.8-5.9) |
| 55-64 | 44 (13.2) | 987 | 4.7 (3.3-6.8) | 2.0 (1.1-3.6) |
| ≥65 | 32 (6.0) | 2059 | 1.7 (1.1-2.7) | 1 [Reference] |
| Sex | ||||
| Male | 62 (50.7) | 1619 | 6.4 (4.8-8.6) | 0.9 (0.7-1.4) |
| Female | 117 (49.3) | 2846 | 5.8 (4.5-7.3) | 1 [Reference] |
| Race/ethnicityc | ||||
| Non-Hispanic white | 147 (88.8) | 3549 | 6.9 (5.6-8.5) | 2.8 (1.4-5.7) |
| Non-Hispanic black | 18 (5.5) | 633 | 2.6 (1.3-5.2) | 1 [Reference] |
| Hispanic | 2 (0.9) | 92 | 1.1 (0.2-4.7) | |
| Non-Hispanic otherd | 5 (2.7) | 110 | 6.1 (1.9-17.8) | 2.4 (0.7-8.4) |
| Multiracial | 7 (2.0) | 81 | 11.9 (4.6-27.1) | 4.2 (1.5-12.2) |
| Education | ||||
| <High school | 23 (13.7) | 502 | 5.1 (2.9-8.8) | 0.9 (0.4-1.9) |
| High school or equivalent | 70 (43.5) | 1579 | 7.5 (5.6-10.0) | 1.4 (0.8-2.5) |
| Some college or technical school with no degree | 61 (34.3) | 1208 | 6.8 (4.9-9.5) | 1.4 (0.8-2.5) |
| College or technical school graduate or higher | 25 (8.5) | 1176 | 2.9 (1.7-4.7) | 1 [Reference] |
| Cigarette smoking status | ||||
| Current smoker (every day) | 92 (56.0) | 577 | 19.4 (15.1-24.6) | 75.1 (24.1-233.8) |
| Current smoker (some days) | 36 (18.4) | 240 | 16.7 (10.7-25.2) | 71.6 (22.0-232.6) |
| Former smoker | 46 (23.8) | 1377 | 5.5 (3.7-8.1) | 28.0 (8.8-89.1) |
| Never smoker | 5 (1.8) | 2271 | 0.2 (0.1-0.7) | 1 [Reference] |
Abbreviations: CI, confidence interval; ENDS, electronic nicotine delivery systems.
aSampling weights of all 5258 records were included in the analysis to account for the complex survey design.
bMultivariate logistic regression model included age, sex, race/ethnicity, education level, and cigarette smoking status. Hosmer-Lemeshow-Satterthwaite goodness-of-fit F test, P = .80.
cNon-Hispanic black and Hispanic respondents were combined to create the reference group.
dIncludes Asian American/Pacific Islander, American Indian/Alaska Native, and any other non-Hispanic racial/ethnic group.
In multivariate analyses, the weighted prevalence of current ENDS use was significantly higher among those who were younger (35-44: PR = 4.2; 95% CI, 2.2-8.0; P < .001) compared with those aged ≥65 and among non-Hispanic white (PR = 2.8; 95% CI, 1.4-5.7) and multiracial (PR = 4.2; 95% CI, 1.5-12.2) participants than among non-Hispanic black or Hispanic participants. We found no difference in prevalence by sex or education. Compared with those who had never smoked cigarettes, the weighted prevalence of current ENDS use was significantly higher for current everyday smokers (PR = 75.1; 95% CI, 24.1-233.8; P < .001), current some days smokers (PR = 71.6; 95% CI, 22.0-232.6; P < .001), and former smokers (PR = 28.0; 95% CI, 8.8-89.1; P < .001; Hosmer-Lemeshow-Satterthwaite goodness-of-fit F test, P = .80; Table 1).
Of the 179 current ENDS users, 95 (weighted prevalence = 60.2%) reported using flavored products. When stratified by age group, the proportion of those using flavored ENDS was highest among those aged 18-24 (9 of 11, weighted prevalence = 75.9%) and lowest among those aged ≥65 (13 of 32, weighted prevalence = 40.1%; test for linear trend = .05).
Smoking Cessation Analyses
Among 809 current cigarette smokers with complete data on the question about using ENDS during the previous year to try to quit smoking cigarettes, 180 (weighted prevalence = 26.1%) answered affirmatively. Current ENDS use was significantly more prevalent among smokers who reported using ENDS to quit smoking (n = 85, weighted prevalence = 43.3%) than among smokers who did not report using ENDS to quit smoking (n = 40, weighted prevalence = 9.9%; PR = 4.4; 95% CI, 2.8-6.8). Those who reported using ENDS to try to quit smoking were significantly more likely to report using ENDS every day (n = 38, weighted prevalence = 20.2%) than those who reported not using ENDS to quit smoking (n = 3, weighted prevalence = 0.4%; PR = 51.9; 95% CI, 14.8-182.2). Although these associations varied by the frequency of cigarette smoking (daily vs some days), the differences were not significant (Breslow-Day test for homogeneity, P = .10).
Of the 1083 current smokers and former smokers who had stopped smoking in the past 5 years, 515 (weighted prevalence = 54.1%; 95% CI, 49.7%-58.4%) reported ever using ENDS. Of those who had used ENDS, 404 (weighted prevalence = 80.3%; 95% CI, 75.0%-84.7%) did not stop smoking (Table 2).
Table 2.
Weighteda odds ratios of 1083 former and current smokers quitting smoking in the past 5 years, by ever having used ENDS, Arkansas Behavioral Risk Factor Surveillance System, 2014
| Ever Used ENDS | Quit (Former Smokers) | Did Not Quit (Current Smokers) | Total |
|---|---|---|---|
| No. (weighted total no.) [%] | |||
| Yes | 111 (65 000) [19.7] | 404 (265 000) [80.3] | 515 (330 000) [100.0] |
| No | 155 (83 000) [29.6] | 413 (197 000) [70.4] | 568 (280 000) [100.0] |
| Total | 266 (148 000) | 817 (462 000) | 1083 (610 000) |
| OR (95% CI) | |||
| Crude weightedb | 0.58 (0.39-0.88) | 1 [Reference] | — |
| Multivariate weightedc | 0.53 (0.34-0.83) | 1 [Reference] | — |
Abbreviations: CI, confidence interval; ENDS, electronic nicotine delivery systems; OR, odds ratio.
aNumbers in thousands are the estimated totals of adults in Arkansas; percentages and ORs are based on these totals. Sampling weights of all 5258 records were included in the analysis to account for the complex survey design.
- Probability of quitting among ENDS users = 65 of 330 (19.7%).
- Probability of not quitting among ENDS users = 265 of 330 (80.3%).
- Odds of quitting among ENDS users = 24.5%.
- Probability of quitting among non-ENDS users = 83 of 280 (29.6%).
- Probability of not quitting| among non-ENDS users: 197 of 280 (70.4%).
- Odds of quitting among non-ENDS users = 42.0%.
- OR = 0.58, or (65 × 197) / (265 × 83) = 0.58.
cMultivariate logistic regression analysis controlled for age, sex, race/ethnicity, and education. Hosmer-Lemeshow-Satterthwaite goodness-of-fit F test, P = .16.
We found that among current smokers and former smokers who had stopped smoking in the past 5 years, having ever used ENDS was significantly associated with reduced odds of quitting smoking (OR = 0.53; 95% CI, 0.34-0.83; Hosmer-Lemeshow-Satterthwaite goodness-of-fit F test, P = .16; Table 2).
Compared with current smokers, any use of ENDS by former smokers was significantly associated with reduced odds that the time since quitting was 1 to 4 years (OR = 0.30; 95% CI, 0.16-0.54; P < .001), but it was not significantly associated with reduced odds that the time since quitting was 6 months to 1 year or <6 months (Table 3).
Table 3.
Weighteda odds ratios of 266 former smokers’ time since quitting smoking, compared with 817 current smokers, by ever having used ENDS, Arkansas Behavioral Risk Factor Surveillance System, 2014
| Ever Used ENDS | Former Smokers’ Time Since Quitting | Total Current Smokers | |||
|---|---|---|---|---|---|
| 1-4 y | 6-11 mo | 3-5 mo | <3 mo | ||
| No. (weighted total No.) [%] | |||||
| Yes | 50 (27 000) [8.3] | 17 (6000) [1.9] | 16 (9000) [2.8] | 28 (22 000) [6.7] | 404 (265 000) [80.3] |
| No | 112 (61 000) [21.6] | 20 (12 000) [4.2] | 7 (3000) [1.1] | 16 (7000) [2.6] | 413 (197 000) [70.4] |
| Total | 162 (88 000) | 37 (18 000) | 23 (12 000) | 44 (29 000) | 817 (463 000) |
| OR (95% CI) | |||||
| Crude weighted | 0.33 (0.19-0.58) | 0.38 (0.16-0.92) | 2.25 (0.64-7.85) | 2.24 (0.89-5.63) | 1 [Reference] |
| Multivariate weightedb | 0.30 (0.16-0.54) | 0.38 (0.14-1.02) | 2.33 (0.63-8.62) | 2.06 (0.77-5.57) | 1 [Reference] |
Abbreviations: CI, confidence interval; ENDS, electronic nicotine delivery systems; OR, odds ratio.
aNumbers in thousands are the estimated totals of adults in Arkansas with the characteristic; percentages and ORs are based on these totals. Sampling weights of all 5258 records were included in the analysis to account for the complex survey design.
bNominal logistic regression analyses controlled for age, sex, race/ethnicity, and education.
Beliefs About the Health Benefits of ENDS
Of the 3808 respondents who ranked their attitude toward the statement “ENDS use is harmful to nonsmokers’ health,” 1712 (weighted prevalence = 41.6%) strongly agreed and 725 (weighted prevalence = 20.9%) agreed with the statement (Table 4). Of the 165 current ENDS users, 33 (weighted prevalence = 17.3%) strongly agreed and 47 (weighted prevalence = 24.0%) agreed with the statement. In multivariate ordinal logistic regression analysis, in which all independent variables were regressed against the different levels of agreement for all respondents, the weighted prevalence of strong disagreement with the statement was significantly higher among men than women (P = .01), among those with less than a high school education than those with other education levels (P < .001), and among current ENDS users than non-current ENDS users (P < .001).
Table 4.
Weighted prevalence of beliefs about harm to health from ENDS use among 3808 adults, by demographic characteristics, education, cigarette smoking status, and current ENDS use status, Arkansas Behavioral Risk Factor Surveillance System, 2014
| Agreement With the Statement “If a Nonsmoker Begins to Use Electronic Cigarettes, It Is Harmful to Their Health,” No. (Weighted %a) | ||||||
|---|---|---|---|---|---|---|
| Variable | Strongly Agree | Somewhat Agree | Neither Agree nor Disagree | Somewhat Disagree | Strongly Disagree | Satterthwaite-Adjusted F Test P Valueb |
| Total respondents | 1712 (41.6) | 725 (20.9) | 638 (18.4) | 311 (9.5) | 422 (9.7) | |
| Age, y | ||||||
| 18-24 | 41 (30.6) | 46 (31.7) | 25 (19.5) | 17 (11.6) | 7 (6.5) | .55 |
| 25-34 | 115 (38.1) | 63 (23.0) | 63 (21.1) | 23 (12.6) | 18 (5.3) | |
| 35-44 | 153 (43.8) | 72 (20.3) | 62 (18.7) | 24 (6.9) | 34 (10.3) | |
| 45-54 | 248 (44.5) | 115 (18.1) | 108 (20.4) | 32 (6.7) | 49 (10.2) | |
| 55-64 | 354 (41.1) | 176 (19.6) | 157 (17.4) | 95 (11.6) | 91 (10.3) | |
| ≥65 | 801 (48.1) | 253 (15.7) | 223 (13.9) | 120 (8.1) | 223 (14.2) | |
| Sex | ||||||
| Male | 571 (38.8) | 280 (20.1) | 258 (20.6) | 123 (9.5) | 162 (10.9) | .01c |
| Female | 1141 (44.2) | 445 (21.6) | 380 (16.3) | 188 (9.4) | 260 (8.5) | |
| Race/ethnicity | ||||||
| Non-Hispanic white | 1385 (41.6) | 593 (21.0) | 532 (19.9) | 239 (9.2) | 272 (8.2) | .35 |
| Non-Hispanic black | 223 (41.1) | 89 (18.6) | 67 (13.5) | 54 (12.6) | 109 (14.3) | |
| Hispanic | 36 (39.0) | 16 (28.6) | 9 (9.8) | 5 (8.3) | 16 (14.3) | |
| Non-Hispanic otherd | 45 (56.5) | 10 (8.1) | 13 (12.1) | 6 (6.9) | 18 (16.4) | |
| Multiracial | 23 (21.3) | 17 (31.0) | 17 (24.5) | 7 (4.9) | 7 (18.4) | |
| Education | ||||||
| <High school | 170 (40.1) | 48 (13.1) | 43 (13.1) | 47 (14.0) | 106 (19.7) | <.001c |
| High school or equivalent | 558 (35.8) | 242 (23.0) | 191 (18.4) | 137 (11.0) | 173 (11.8) | |
| Some college or technical school with no degree | 485 (44.9) | 199 (19.7) | 206 (20.6) | 78 (9.0) | 85 (5.8) | |
| College or technical school graduate or higher | 499 (48.1) | 236 (25.8) | 198 (19.3) | 49 (3.5) | 58 (3.4) | |
| Cigarette smoking status | ||||||
| Current smoker (every day) | 162 (35.5) | 102 (17.2) | 111 (21.8) | 74 (16.5) | 52 (9.2) | .73 |
| Current smoker (some days) | 66 (31.3) | 53 (30.3) | 43 (18.1) | 18 (9.1) | 24 (11.2) | |
| Former smoker | 520 (41.7) | 229 (21.9) | 184 (16.3) | 85 (7.6) | 148 (12.6) | |
| Never smoker | 964 (45.0) | 341 (20.5) | 300 (18.3) | 134 (8.0) | 198 (8.1) | |
| Current ENDS use status | ||||||
| Yes | 33 (17.3) | 47 (24.0) | 32 (22.2) | 23 (17.2) | 30 (19.4) | <.001c |
| No | 1679 (43.2) | 678 (20.7) | 606 (18.1) | 288 (9.0) | 392 (9.0) | |
Abbreviation: ENDS, electronic nicotine delivery systems.
aSampling weights of all 5258 records were included in the analysis to account for the complex survey design.
bMultivariate ordinal logistic regression models included age, sex, race/ethnicity, and education.
cSignificant at P < .05.
dIncludes Asian American/Pacific Islander, American Indian/Alaska Native, and any other non-Hispanic racial/ethnic group.
Of the 3658 respondents to the statement “Switching to ENDS does not reduce the tobacco-related health risk of current cigarette smokers,” 1000 (weighted prevalence = 25.1%) strongly agreed and 793 (weighted prevalence = 21.9%) agreed with the statement. Of the 168 current ENDS users, 14 (weighted prevalence = 9.4%) strongly agreed and 36 (weighted prevalence = 24.5%) agreed with the statement. For all respondents, the weighted prevalence of strong disagreement with this statement was significantly higher among men than women (P = .03) and among current ENDS users than noncurrent ENDS users (P < .001; Table 5).
Table 5.
Weighted prevalence of beliefs about reduction of tobacco-related health risks of current smokers who switched to ENDS among 3658 adults, by demographic characteristics, education, cigarette smoking status, and current ENDS use status, Arkansas Behavioral Risk Factor Surveillance System, 2014
| Agreement With the Statement “If a Current Cigarette Smoker Switches to Electronic Cigarettes, It Does Not Reduce Their Tobacco-Related Health Risks,” No. (Weighted %a) | ||||||
|---|---|---|---|---|---|---|
| Variable | Strongly Agree | Somewhat Agree | Neither Agree nor Disagree | Somewhat Disagree | Strongly Disagree | Satterthwaite-Adjusted F-Test P Valueb |
| All respondents | 1000 (25.1) | 793 (21.9) | 620 (18.4) | 627 (17.9) | 618 (16.8) | |
| Age, y | ||||||
| 18-24 | 28 (21.1) | 38 (26.0) | 23 (20.5) | 24 (20.3) | 18 (12.1) | .55 |
| 25-34 | 60 (20.7) | 62 (22.0) | 63 (22.9) | 57 (19.3) | 32 (15.2) | |
| 35-44 | 83 (25.4) | 79 (22.3) | 55 (17.2) | 57 (15.1) | 64 (20.0) | |
| 45-54 | 136 (24.2) | 115 (20.2) | 98 (17.8) | 112 (20.8) | 88 (17.1) | |
| 55-64 | 212 (25.1) | 207 (23.3) | 139 (17.7) | 169 (19.3) | 118 (14.7) | |
| ≥65 | 481 (32.6) | 292 (19.2) | 242 (14.8) | 208 (13.4) | 298 (19.9) | |
| Sex | ||||||
| Male | 347 (22.3) | 292 (22.0) | 243 (19.7) | 236 (17.8) | 238 (18.2) | .03c |
| Female | 653 (27.7) | 501 (21.9) | 377 (17.1) | 391 (17.9) | 380 (15.4) | |
| Race/ethnicity | ||||||
| Non-Hispanic white | 786 (23.5) | 653 (22.0) | 515 (20.0) | 513 (19.3) | 421 (15.2) | .62 |
| Non-Hispanic black | 145 (26.2) | 93 (23.8) | 74 (13.2) | 86 (14.5) | 139 (22.3) | |
| Hispanic | 27 (40.2) | 13 (12.9) | 12 (14.5) | 8 (11.6) | 20 (20.8) | |
| Non-Hispanic otherd | 27 (36.8) | 13 (23.3) | 8 (5.5) | 12 (8.5) | 26 (25.9) | |
| Multiracial | 15 (16.3) | 21 (38.1) | 11 (14.4) | 8 (9.9) | 12 (21.4) | |
| Education | ||||||
| <High school | 123 (36.3) | 56 (15.0) | 43 (15.0) | 48 (9.0) | 120 (24.8) | .61 |
| High school or equivalent | 344 (22.4) | 277 (25.1) | 184 (15.7) | 220 (18.6) | 244 (18.2) | |
| Some college or technical school with no degree | 282 (24.4) | 219 (19.7) | 184 (20.4) | 191 (21.4) | 141 (14.1) | |
| College or technical school graduate or higher | 251 (21.5) | 241 (25.8) | 209 (22.9) | 168 (18.1) | 113 (11.7) | |
| Cigarette smoking status | ||||||
| Current smoker (every day) | 106 (21.8) | 136 (26.2) | 86 (18.0) | 94 (17.1) | 86 (17.0) | .30 |
| Current smoker (some days) | 46 (21.3) | 46 (28.3) | 36 (15.1) | 44 (21.5) | 35 (13.8) | |
| Former smoker | 318 (26.1) | 238 (19.3) | 174 (15.8) | 190 (17.1) | 210 (21.7) | |
| Never smoker | 530 (26.3) | 373 (20.8) | 324 (20.4) | 299 (18.1) | 287 (14.4) | |
| Current ENDS use | ||||||
| Yes | 14 (9.4) | 36 (24.5) | 18 (8.3) | 44 (23.4) | 56 (34.3) | <.001c |
| No | 986 (26.2) | 757 (21.8) | 602 (19.1) | 583 (17.5) | 562 (15.5) | |
Abbreviation: ENDS, electronic nicotine delivery systems.
aSampling weights of all 5258 records were included in the analytic results to account for the complex survey design.
bMultivariate ordinal logistic regression analyses controlled for age, sex, race/ethnicity, and education.
cSignificant at P < .05.
dIncludes Asian American/Pacific Islander, American Indian/Alaska Native, and any other non-Hispanic racial/ethnic group.
Smoking Prevalence
We found no significant changes in the overall prevalence of current and former cigarette smoking in Arkansas from 2011 to 2014 using a weighted age-adjusted analysis (Wald F test, P = .50 for current smoking, P = .80 for former smoking). The estimated annual percentage change in prevalence during this period was −2.8% (95% CI, −12.9% to 8.4%, P = .40) for smoking every day, 0.9% (95% CI, −7.6% to 10.3%, P = .70) for smoking some days, and 0.6% (95% CI, −4.1% to 5.7%, P = .60) for former smoking (Figure).
Figure.
Prevalence of current and former cigarette smoking among adults, Arkansas Behavioral Risk Factor Surveillance System, 1996-2014. Beginning in 2011, the system included both landline and cellular telephones in the sampling frame and used a new method for weighting. As such, estimates for 2011-2014 are not strictly comparable with estimates from previous years.
Discussion
Our most important finding was that having ever used ENDS was significantly associated with reduced odds of quitting smoking. When controlled for age, sex, race/ethnicity, and education, the respondents who had ever used ENDS reduced their chances of successfully quitting smoking by about 50%. Furthermore, the finding that >80% of those who indicated ever using ENDS did not quit smoking suggests that ENDS use may actually promote nicotine addiction and result in users simply adding ENDS use to cigarette smoking, without ever switching from cigarette smoking to ENDS. On a related note, any use of ENDS by former smokers significantly reduced their odds that the time since quitting was 1 to 4 years. These reduced odds were noted for the time since quitting of 6 to 12 months, but the reduction was not significant. This reduction in the odds of quitting was not noted for the time since quitting of <6 months, possibly because of the relatively small number of respondents in this group or because of the high relapse rates among smokers found in most trials.20 As such, our study showed a significant inverse association between ENDS use and the odds of successfully stopping smoking (for >1 year). These findings are consistent with a study that used data from the National Youth Tobacco Survey, which found an association between ENDS use and a reduced likelihood of cigarette smoking cessation.21 Our results are also consistent with the results of a combined systematic review and meta-analysis, which found that the odds of quitting cigarette smoking were 28% lower in those who had ever used ENDS compared with those who did not use ENDS.13
We also found that use of ENDS was highest among cigarette smokers in our study who had tried to quit. In addition, when we looked at Arkansas BRFSS statistics from 2011 to 2014, the period during which ENDS use was gaining traction in the market, we found no significant change in the prevalence of current or former cigarette smoking. These findings, along with the aforementioned results, suggest that ENDS may not be effective tools for smoking cessation. In fact, ENDS use may actually promote nicotine addiction.
In our study, the prevalence of current ENDS use in Arkansas was 6.1%; of ever using ENDS, 20.6%; and of current ENDS users also being current cigarette smokers, 74.4%. These results are similar to the results of a 2013 Kansas Adult Tobacco Survey of 9656 adult respondents, which reported that 14% of adult cigarette smokers were current ENDS users, 45% had tried ENDS at some time, and 77% of current ENDS users were also current smokers.5 The relatively high prevalence of ENDS use in Arkansas and Kansas, both of which have large rural populations, suggests that ENDS use has penetrated into rural areas of the United States.
Our assessment of beliefs about ENDS held by the general Arkansas population and current ENDS users found that many current ENDS users believed that ENDS were not harmful to the health of nonsmokers and that switching from smoking to ENDS would reduce tobacco-related risks for current smokers. These beliefs differed from those of most of our study population, which thought that ENDS use was likely to be harmful to the health of nonsmokers and that switching from cigarettes to ENDS was not likely to reduce the tobacco-related health risks of smokers. At least 1 major review indicated that the safety of ENDS use is questionable.14 These possibly misinformed beliefs of ENDS users in our study suggest an opportunity for public health education focused on explaining the risks of ENDS use for nonsmokers and smokers.
We also assessed smokers’ beliefs about ENDS and smoking cessation. We found that ENDS use was highest among cigarette smokers who had tried to quit smoking, suggesting that many smokers in Arkansas believed that ENDS use might help them quit smoking. This finding is consistent with other studies that showed that many smokers choose ENDS as an alternative to other US Food and Drug Administration–approved smoking cessation aids.6,22 These findings suggest the importance of telling cigarette smokers that ENDS use may not help them quit smoking.
Our study also found that the state-specific ENDS-related belief questions that were added to the Arkansas BRFSS in 2014 provided valuable data on the opinions of the general population, who seemed to be generally aware of the risks of ENDS use, and the beliefs of smokers and ENDS users, who seemed to be less aware of these risks. We encourage other states to add this optional module to the BRFSS because the data obtained may provide their public health educators and policy makers with a more complete understanding of the beliefs about ENDS use and smoking in their states.
Limitations
Our study had several limitations. First, the 2014 BRFSS survey did not collect data on the quantity of cigarettes smoked or ENDS used. Stratifying the data by the different levels of cigarette use and ENDS use would have provided additional information about the degree to which nicotine addiction could have affected the results. Second, we assumed that ENDS use began after, rather than before, cigarette smoking, but the BRFSS questionnaire did not include questions on when ENDS use began. Answers to the questions about beliefs could have been affected by whether or not participants had already failed to quit smoking by using ENDS. Third, we excluded 793 of the 5258 survey participants (15.1%) because of incomplete responses on our variables of interest, and we excluded additional participants from the beliefs questions because of incomplete responses or nonresponse, which may have introduced bias. However, our sensitivity analyses indicated that our results did not change substantially when we excluded nonrespondents. Finally, although the cross-sectional design of our study allowed us to determine associations between variables, it restricted our ability to draw definitive causal inferences, particularly about the association between ENDS use and smoking cessation. Nevertheless, the association between ENDS use and attempts at smoking cessation suggests that a substantial proportion of smokers believe that ENDS use will help with smoking cessation. Furthermore, the inverse association between ENDS use and smoking cessation suggests that ENDS use may actually lower the likelihood of smoking cessation.
Conclusions
In our study, ENDS use reduced the odds of quitting smoking by almost half and was inversely associated with smoking cessation. Most smokers who tried ENDS did not stop smoking. Only a small proportion of current ENDS users believed that ENDS was harmful to the health of nonsmokers or that switching from smoking to ENDS use did not reduce the tobacco-related health risks of smokers. Tobacco cessation programs should convey the message to cigarette smokers that ENDS use may not help smokers quit smoking and that it may actually reduce their likelihood of quitting smoking. These findings can inform public health policy making and educational efforts at the state and regional levels.
Acknowledgments
The Arkansas Department of Health designs and operates the BRFSS in collaboration with the Centers for Disease Control and Prevention. We thank Monica Whitwell, MEd, BRFSS coordinator, for making the 2014 Arkansas BRFSS data available to us.
Footnotes
Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.
References
- 1. Franck C, Budlovsky T, Windle SB, Filion KB, Eisenberg MJ. Electronic cigarettes in North America: history, use, and implications for smoking cessation. Circulation. 2014;129(19):1945–1952. [DOI] [PubMed] [Google Scholar]
- 2. Kann L, McManus T, Harris WA, et al. Youth Risk Behavior Surveillance—United States, 2015 [published erratum appears in MMWR Morb Mortal Wkly Rep. 2016;65(23):610]. MMWR Surveill Summ. 2016;65(6):1–174. [DOI] [PubMed] [Google Scholar]
- 3. Schoenborn CA, Gindi RM. Electronic Cigarette Use Among Adults: United States, 2014. Hyattsville, MD: National Center for Health Statistics; 2015. NCHS data brief 217. [PubMed] [Google Scholar]
- 4. Agaku IT, King BA, Husten CG, et al. Tobacco product use among adults—United States, 2012-2013 [published erratum appears in MMWR Morb Mortal Wkly Rep. 2014;63(26):576]. MMWR Morb Mortal Wkly Rep. 2014;63(25):542–547. [PMC free article] [PubMed] [Google Scholar]
- 5. Christensen T, Welsh E, Faseru B. Profile of e-cigarette use and its relationship with cigarette quit attempts and abstinence in Kansas adults. Prev Med. 2014;69:90–94. [DOI] [PubMed] [Google Scholar]
- 6. Pokhrel P, Fagan P, Little MA, Kawamoto CT, Herzog TA. Smokers who try e-cigarettes to quit smoking: findings from a multiethnic study in Hawaii. Am J Public Health. 2013;103(9):e57–e62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Mokdad AH. The Behavioral Risk Factors Surveillance System: past, present and future. Annu Rev Public Health. 2009;30:43–54. [DOI] [PubMed] [Google Scholar]
- 8. Jamal A, Thun M, Yu XQ, et al. Changes in smoking prevalence among US adults by state and region: estimates from the Tobacco Use Supplement to the Current Population Survey, 1992-2007. BMC Public Health. 2011;11:512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Jamal A, Agaku IT, O’Connor E, King BA, Kenemer JB, Neff L. Current cigarette smoking among adults—United States, 2005-2013. MMWR Morb Mortal Wkly Rep. 2014;63(47):1108–1112. [PMC free article] [PubMed] [Google Scholar]
- 10. US Centers for Disease Control and Prevention. BRFSS prevalence and trends data. http://www.cdc.gov/brfss/brfssprevalence/index.html. Accessed March 16, 2016.
- 11. Grana RA, Ling PM. “Smoking revolution”: a content analysis of electronic cigarette retail websites. Am J Prev Med. 2014;46(4):395–403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Quit Smoking Community. Best electronic cigarette brands for 2016. https://quitsmokingcommunity.org/electronic-cigarettes. Accessed August 26, 2016.
- 13. 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):116–128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Pisinger C, Døssing M. A systematic review of health effects of electronic cigarettes. Prev Med. 2014;69:248–260. [DOI] [PubMed] [Google Scholar]
- 15. Siu AL; US Preventive Services Task Force. Behavioral and pharmacotherapy interventions for tobacco smoking cessation in adults, including pregnant women: US Preventive Services Task Force recommendation statement. Ann Intern Med. 2015;163(8):622–634. [DOI] [PubMed] [Google Scholar]
- 16. Arkansas Department of Health. What is the Behavioral Risk Factor Surveillance System? http://www.healthy.arkansas.gov/programsServices/healthStatistics/Brfss/Pages/default.aspx. Accessed August 4, 2016.
- 17. McCullagh P. Regression models for ordinal data. J R Stat Soc. 1980;42:109–142. http://www.stat.uchicago.edu/∼pmcc/pubs/paper2.pdf. Accessed March 16, 2016. [Google Scholar]
- 18. Research Triangle Institute. SUDAAN Language Manual, Volumes 1 and 2, Release 11. Research Triangle Park, NC: Research Triangle Institute; 2012. [Google Scholar]
- 19. National Cancer Institute. Statistical Research and Applications Branch, Joinpoint Regression Program, version 4.0.4. http://surveillance.cancer.gov/joinpoint. Published May 2013. Accessed March 16, 2016.
- 20. Patnode CP, Henderson JT, Thompson JH, Senger CA, Fortmann SP, Whitlock EP. Behavioral Counseling and Pharmacotherapy Interventions for Tobacco Cessation in Adults, Including Pregnant Women: A Review of Reviews for the US Preventive Services Task Force. Rockville, MD: Agency for Healthcare Research and Quality; 2015. Evidence synthesis 134. AHRQ publication 14-05200-EF-1. [PubMed] [Google Scholar]
- 21. Dutra LM, Glantz SA. Electronic cigarettes and conventional cigarette use among US adolescents: a cross-sectional study [published erratum appears in JAMA Pediatr. 2014;168(7):684]. JAMA Pediatr. 2014;168(7):610–617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Rutten LJ, Blake KD, Agunwamba AA, et al. Use of e-cigarettes among current smokers: associations among reasons for use, quit intentions, and current tobacco use. Nicotine Tob Res. 2015;17(10):1228–1234. [DOI] [PMC free article] [PubMed] [Google Scholar]