Drug and Alcohol Dependence 149 (2015) 25–30
Contents lists available at ScienceDirect
Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep
Risky behaviors, e-cigarette use and susceptibility of use among college students M.L. Saddleson a,∗ , L.T. Kozlowski a , G.A. Giovino a , L.W. Hawk b , J.M. Murphy c , M.G. MacLean d , M.L. Goniewicz f , G.G. Homish a , B.H. Wrotniak g , M.C. Mahoney a,e a University at Buffalo, State University of New York, School of Public Health and Health Professions, Department of Community Health and Health Behavior, Buffalo, NY, USA b University at Buffalo, State University of New York; Department of Psychology, Buffalo, NY, USA c State University of New York, at Cortland, Health Department, Cortland, NY, USA d State University of New York, Buffalo State, Department of Psychology, Buffalo, NY, USA e Roswell Park Cancer Institute, Department of Medicine and Department of Health Behavior, Buffalo, NY, USA f Roswell Park Cancer Institute, Division of Cancer Prevention and Population Sciences, Buffalo, NY, USA g D’Youville College, Center for Health Behavior Research, Buffalo, NY, USA
a r t i c l e
i n f o
Article history: Received 11 September 2014 Received in revised form 11 December 2014 Accepted 2 January 2015 Available online 24 January 2015 Keywords: Electronic cigarettes Tobacco Smoking Susceptibility to smoking Young adults College students
a b s t r a c t Background: Since 2007, there has been a rise in the use of electronic cigarettes (e-cigarettes). The present study uses cross-sectional data (2013) to examine prevalence, correlates and susceptibility to e-cigarettes among young adults. Methods: Data were collected using an Internet survey from a convenience sample of 1437, 18–23 year olds attending four colleges/universities in Upstate New York. Results were summarized using descriptive statistics; logistic regression models were analyzed to identify correlates of e-cigarette use and susceptibility to using e-cigarettes. Results: Nearly all respondents (95.5%) reported awareness of e-cigarettes; 29.9% were ever users and 14.9% were current users. Younger students, males, non-Hispanic Whites, respondents reporting average/below average school ability, ever smokers and experimenters of tobacco cigarettes, and those with lower perceptions of harm regarding e-cigarettes demonstrated higher odds of ever use or current use. Risky behaviors (i.e., tobacco, marijuana or alcohol use) were associated with using e-cigarettes. Among never e-cigarette users, individuals involved in risky behaviors or, with lower harm perceptions for e-cigarettes, were more susceptible to future e-cigarette use. Conclusions: More e-cigarette users report use of another nicotine product besides e-cigarettes as the first nicotine product used; this should be considered when examining whether e-cigarette use is related to cigarette susceptibility. Involvement in risky behaviors is related to e-cigarette use and susceptibility to e-cigarette use. Among college students, e-cigarette use is more likely to occur in those who have also used other tobacco products, marijuana, and/or alcohol. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Since 2007, electronic cigarettes (e-cigarettes) have gained popularity in the U.S. In April 2014, the Food and Drug Administration (FDA) proposed regulations restricting e-cigarette sales to minors. One public health concern is that e-cigarettes may be a “gateway,” leading e-cigarette users to begin smoking (Cobb and Abrams, 2011). E-cigarettes are often believed by users to be less
∗ Corresponding author at: 3435 Main St., Buffalo, NY 14214, USA. Tel.: +1 716 829 5702. E-mail address: [email protected] (M.L. Saddleson). http://dx.doi.org/10.1016/j.drugalcdep.2015.01.001 0376-8716/© 2015 Elsevier Ireland Ltd. All rights reserved.
harmful than cigarettes (Choi and Forster, 2013; Pearson et al., 2012). A recent review supports beliefs about reduced harmfulness, concluding “Health professionals may consider advising smokers unable or unwilling to quit through other routes to switch to e-cigarettes as a safer alternative to smoking and a possible pathway to complete cessation of nicotine use” (Hajek et al., 2014). Perceptions of e-cigarettes as less harmful than cigarettes may be associated with use of e-cigarettes. Perceived harmfulness of marijuana is strongly associated with marijuana use (Bachman et al., 1998; Bailey et al., 1992). Existing studies offer mixed results on associations between e-cigarette harm perceptions and use (Adkison et al., 2013; Choi and Forster, 2013; Sutfin et al., 2013).
26
M.L. Saddleson et al. / Drug and Alcohol Dependence 149 (2015) 25–30
Prevalence of e-cigarette ever use is highest among young adults (Adkison et al., 2013; Pearson et al., 2012; Regan et al., 2013). Young adults are more aware and have higher rates of use than older groups (Choi and Forster, 2013; Pearson et al., 2012; Pepper et al., 2013; Sutfin et al., 2013; Trumbo and Harper, 2013). Trumbo et al. surveyed students at a Colorado university in 2011: 13.0% had ever used e-cigarettes (Trumbo and Harper, 2013). Sutfin et al. surveyed undergraduates in North Carolina colleges in 2009 (in early years of e-cigarette marketing), reporting prevalence of 4.9%; those who were smokers, male, Hispanic, “other race,” or in Greek organizations (sorority/fraternity), or with lower harm perceptions of e-cigarettes were likelier to use e-cigarettes (Sutfin et al., 2013). Latimer et al. surveyed urban public universities in Texas during 2011 and reported 3.1% used e-cigarettes during the previous 30days (Latimer et al., 2013). In 2014, e-cigarette prevalence rates among middle and high school students were 6.5% (ever use) and 2.0% (current use; Dutra and Glantz, 2014). The present study measured prevalence and correlates of e-cigarette use among college students attending four colleges/universities in Upstate New York in 2013. Given concerns about e-cigarette use in never smokers of cigarettes, analyses explored correlates of e-cigarette use in never smokers. We examined factors that could lead young adults to try e-cigarettes. Among never e-cigarette users, we assessed correlates of “susceptibility” to future e-cigarette use adapting measures Pierce (Pierce et al., 1996, 2005, 1995) employed to examine ‘susceptibility to cigarette smoking’ which predicted never smokers who later become cigarette smokers (Pierce et al., 1996; Zhu et al., 2013). 2. Methods 2.1. Sample Undergraduate students in selected classes (e.g., psychology/health behaviorrelated courses) at four colleges/universities in New York State (NYS; outside of New York City) participated in fall, 2013, providing informed consent for this IRBapproved research. 2.2. Survey instrument Our 111-item, self-administered, web-based survey, used items from published studies on e-cigarettes, assessed awareness, ever and current use (i.e., within the past 30-days). Skip patterns determined items completed. We collected students’ individual school identification number and school email address, to give credit for participation and to ensure surveys were completed only once. The survey assessed demographic characteristics and other health risk behaviors, such as, cigarette use, use of other nicotine and tobacco products, patterns of alcohol use, and marijuana use. 2.3. Procedures A Web-based survey was used. At one college/university (n = 875), the survey was available through a psychology research website that was only accessible to introductory psychology students (PSY 101). At the other colleges/universities, students (n = 137, n = 81, and n = 333, respectively; 11 students did not provide university attendance information) accessed the survey website as directed by instructors. At the discretion of the investigator(s) at each college/university, respondents were granted either some form of course credit/research credits or entered into a lottery for a $25 grocery store gift card. To accurately represent a college population and because the legal age to provide consent is age 18 years, students younger than 18 years old and older than 23 were excluded from analyses (n = 111). International students were excluded from analyses (n = 100) due to cultural values which are distinct from U.S. college students and the variable availability of e-cigarettes elsewhere. 2.4. Measures 2.4.1. Demographic characteristics. Age was recoded as 18, 19 and 20–23 years to keep distributions similar, yet illustrate multiple age groups. 2.4.2. Gender. Male or female. 2.4.3. Race/ethnicity. Created from two separate items (Race and Ethnicity: “Are you Spanish, Hispanic or Latino?”), a 3 category variable was constructed: Non-Hispanic
White/Caucasian, Non-Hispanic non-whites (all races excluding white/Caucasian), and Hispanic (regardless of race). 2.4.4. School ability. Assessed by: “How well do you do in school? Would you say. . .,” “Much better than average,” “Better than average”, “Average,” “Below average” or, “Don’t know.” “Much better than average” and “better than average” were combined and then recoded into 2 categories: “Average/below average/don’t know” and “better than average”. 2.4.5. Awareness of e-cigarettes. One item asked about awareness of e-cigarettes: “Prior to today, have you ever heard about electronic cigarettes (e-cigarettes)?” Responses included: “No, I have never read anything about them and have never been told about them;” “Yes, I read a bit about them or someone told me about them;” “Yes, I am informed on the e-cig, but I have never used it;” “Yes, and I have already used an e-cigarette.” A binary variable was created (yes/no); any “yes” response indicated awareness. 2.4.6. E-cigarette ever use. Assessed by: “Have you ever tried or experimented with an e-cigarette, even one or two puffs?” Those who responded “yes” were classified as ever users. 2.4.7. E-cigarette use. Current use included use on one or more days in the previous 30 days. Respondents who ever used e-cigarettes, but not in the previous 30 days, were classified as discontinued e-cigarette users. 2.4.8. Tobacco cigarette smoking status. Based on having ever tried or experimented with tobacco cigarettes, the number of days smoked in the past 30 days, and the number of cigarettes smoked in one’s lifetime, a four category variable was created: Never smokers (never tried a tobacco cigarette), former smokers (smoked ≥ 100 cigarettes in lifetime, and have smoked 0 out of the past 30 days), experimenters (have ever tried a cigarette, have smoked < 100 cigarettes in lifetime, and have smoked 0 of the past 30 days), and current smokers (have smoked at least 1 day out of the past 30). For the multivariable analyses, smoking status was collapsed into three categories (never smokers, experimenters and ever [current and former smokers]), due to few former smokers in our sample (n = 17). 2.4.9. Other tobacco use. Assessed by one item: “From the following list, please check any of the tobacco products, besides cigarettes, you have used in the last 30 days:” (responses included: cigars, pipes, chewing tobacco, snuff, snus, hookah, clove cigarettes, bidis, other, or I have not used any other tobacco products). A binary variable was created (any/none). 2.4.10. First nicotine product used. A two-category variable was created from: Which one of the following was the first nicotine product that you used? (same responses as above 2.4.9). A three category variable was created: never used a nicotine product, first used a form of nicotine besides e-cigarettes, and first used e-cigarettes. 2.4.11. Alcohol use. Two items, (1) “During the past 30 days, how many days did you have at least one drink of any alcoholic beverage?” Responses were scored as binary (any/none), “Considering all types of alcoholic beverages, how many times during the past 30 days did you have 5 (for males)/4 (for females) or more drinks on an occasion?” Responses were recoded (any/none). 2.4.12. Marijuana use. One item, “During the last 12 months, how often did you use marijuana (cannabis, weed, pot)?” was scored as a binary variable (any/none). 2.4.13. Perceived harm of e-cigarettes. Respondents chose from a 5-category Likert scale, ranging from strongly agree to strongly disagree: “E-cigarettes are less harmful than tobacco cigarettes.” “Strongly agree” and “agree” were combined to form the “agree” group, and “strongly disagree” and “disagree” into “disagree;” a twocategory variable was then created to distinguish those who think e-cigarette use is dangerous (disagreed that e-cigarettes were less harmful than tobacco cigarettes) from those who think it is not dangerous (agreed/neutral about the harm of ecigarettes): disagree versus agree/neutral. 2.4.14. Susceptibility to tobacco cigarette smoking. This measure was adapted from Pierce’s susceptibility measure and was assessed in never users of cigarettes (Pierce et al., 1996, 2005, 1995). The measure combined two questions: “Do you think that you will try cigarettes soon?” and “If one of your best friends were to offer you a cigarette, would you try it?” Responses included: definitely yes, probably yes, probably not, definitely not and don’t know. A binary variable was created; any response other than “definitely not” qualified respondents as susceptible to cigarette smoking. 2.4.15. Susceptibility to e-cigarette use. This employed two items similar to cigarette susceptibility, except related to e-cigarettes. A binary variable was created (susceptible/not susceptible). This susceptibility measure is used to predict future e-cigarette use among those who have never tried e-cigarettes.
M.L. Saddleson et al. / Drug and Alcohol Dependence 149 (2015) 25–30 A refined e-cigarette susceptibility measure was adapted from our original measure classifying respondents as susceptible only if they, at minimum, said “probably yes” to the two susceptibility questions (excluding those responding probably not, definitely not or, don’t know); these respondents are categorized as “highly susceptible.” The distribution of ‘susceptibility to e-cigarettes’ was based on 5 categories of responses (definitely yes, probably yes, probably not, definitely not, don’t know) to two questions: (1) “Do you think you will try an e-cigarette soon?” and, (2) “If one of your best friends were to offer you an e-cigarette, would you try it?” When combined, using the traditional susceptibility definition (Pierce et al., 1995, 1996, 2005), 39.2% of never e-cigarette users are susceptible. Including only those with higher susceptibility (responded definitely yes and/or probably yes), our alternate measure estimated 10.1% of never e-cigarette users as highly susceptible. 2.5. Statistical analysis Descriptive statistics were used to summarize results (Stata version 13, Stata Corporation, College Park, TX). Chi-square tests assessed differences between institutions, e-cigarette ever versus never users, and e-cigarette current versus not current users. Multivariate logistic regression models evaluated the independent influence of demographic variables, selected health behavior and belief variables on ever use of e-cigarettes and concurrent use of e-cigarettes smoked cigarettes. These same independent variables were used to predict susceptibility to both use of tobacco cigarettes and e-cigarettes. Adjusted odds ratios (AOR) and 95% confidence intervals (CI) were computed; two-sided alpha level was 0.05. Multivariable models controlled for demographic variables (age, gender, and race/ethnicity) and other variables represented in the models.
3. Results Respondents included 1437 students (18–23 years) from four colleges/universities. See Table 1. There are differences between colleges/universities for age, gender, race/ethnicity (p < 0.001) and cigarette smoking status (p = 0.001). Overall, 95.5% of respondents report awareness of e-cigarettes; 29.9% are ever users and 14.9% are current users of e-cigarettes, with 6.4% reporting concurrent use of both e-cigarettes and tobacco cigarettes. 3.1. Correlates of ever use of e-cigarettes Table 2 presents descriptive statistics and multivariable analyses for both ever use and current use of e-cigarettes. Lower odds of ever use of e-cigarettes are seen among persons 20–23 years old, females, non-Hispanic non-whites, and those reporting better than average school performance. Those who have experimented with cigarette smoking have a 5-fold greater odds of having ever used e-cigarettes (AOR = 5.52, 95% CI = 3.89–7.85) and ever smokers of cigarettes have nearly 20 times greater odds of having ever used e-cigarettes (AOR = 19.10, 95% CI = 12.12–30.12). Ever use of ecigarettes is also higher among those who used any other tobacco products beside cigarettes in the past 30 days (AOR = 2.38, 95% CI = 1.72–3.29), those who used marijuana in the past 12 months (AOR = 2.37, 95% CI = 1.69–3.31) and those with lower e-cigarette harm perceptions (AOR = 1.78, 95% CI = 1.10–2.86). 3.2. Correlates of current use of e-cigarettes Lower odds of current e-cigarette use are seen in persons 20–23 years old, females, non-Hispanic non-whites, and those reporting better than average school performance. Those who experimented with cigarettes (AOR = 3.12, 95% CI = 1.99–4.91) or those who were ever cigarette smokers (former and current smokers) (AOR = 6.60, 95% CI = 4.09–10.63) have greater odds of current e-cigarette. Current e-cigarette use is also more likely among students who reported any binge drinking in the past 30 days (AOR = 1.71, 95% CI = 1.01–2.90), used tobacco products other than cigarettes during the past 30 days (AOR = 3.11, 95% CI = 2.13–4.54), and have lower perceptions about the harms of e-cigarettes (AOR = 1.34, 95% CI = 1.01–2.90).
27
To explore correlates of e-cigarette use among persons with relatively little cigarette experience, multivariable models assessed e-cigarette ever use and current use among a sub-sample of those who have smoked less than 100 tobacco cigarettes in their lifetime (n = 1257), which included never smokers (64.7%), experimenters (26.8%), and current smokers (8.5%) (those reporting smoking < 100 cigarettes ever, but had smoked within the past 30 days). Results among this subsample are similar to those for the entire sample, suggesting that our findings primarily apply to those who are not yet established tobacco cigarette smokers. However, one notable difference among this subgroup (for both ever and current users of e-cigarettes), is a significant effect for marijuana use (ever use AOR = 3.15, 95% CI = 2.30–4.33; current use AOR = 1.90; 95% CI = 1.26–2.85). 3.3. Susceptibility to tobacco cigarette smoking A logistic regression model was examined among never cigarette smokers (n = 862) to determine whether e-cigarette use was correlated with susceptibility to initiating smoking of tobacco cigarettes, controlling for demographics and first nicotine product used. Among never smokers (n = 862), in terms of first nicotine product used, 70.0% had never used a nicotine product, 27.7% had used some other form besides e-cigarettes first. The highest reported usage was hookah (14.4%), cigars followed (9.7%), and 2.3% reported using an e-cigarette first (other products had rates ≤2.0%). The majority of ever smokers used some other form of nicotine besides e-cigarettes as their first product, with only 2.5% having used e-cigarettes first. Our findings indicate that e-cigarette use is not related to susceptibility to smoking cigarettes, rather, e-cigarette users more commonly first used another nicotine product (86.9%) compared with first using e-cigarettes (7.2%) (p < 0.001) [5.9% had never used any nicotine product]. 3.4. Susceptibility to e-cigarette use Table 3 examines susceptibility to e-cigarette use among never users of e-cigarettes. Never e-cigarette users who were younger (18 or 19 years old) are more likely to be susceptible to e-cigarette use, as are those who drank any amount of alcohol in the past 30 days or, reported any past-month binge drinking. Those who used other tobacco products besides cigarettes in the past 30 days, those who had experimented with smoking tobacco cigarettes or, were ever tobacco cigarette smokers, and those who had lower perceptions about the harms of e-cigarettes were each more likely to be susceptible to e-cigarette use. Using the alternate susceptibility measure to evaluate higher levels of susceptibility to e-cigarette use, never e-cigarette users who identified themselves as non-Hispanic other race (AOR = 2.20, 95% CI = 1.23–3.95), those who used marijuana in the past 12 months (AOR = 2.63, 95% CI = 1.49–4.62), were ever cigarette smokers (AOR = 15.44, 95% CI = 6.91–34.50), had ever experimented with cigarettes (AOR = 3.12, 95% CI = 1.79–5.45), or current users of other tobacco products (AOR = 1.74, 95% CI = 1.01–2.99) have greater odds of being susceptible to using e-cigarettes. As found previously, males and current or former cigarette smokers (18–23 years old) are more likely to ever use e-cigarettes (Adkison et al., 2013; Choi and Forster, 2013; Goniewicz and Zielinska-Danch, 2012; Regan et al., 2013). Results on race/ethnicity are mixed (Choi and Forster, 2013; Pearson et al., 2012; Regan et al., 2013; Sutfin et al., 2013). Sutfin et al. (2013) found Hispanics and those of “other” races, compared to Whites, to be about twice as likely to have used e-cigarettes; while we found those of nonHispanic “other” (non-White) races are one-half as likely to have ever used e-cigarettes compared to Whites. These differences could relate to yet unidentified geographic differences.
28
M.L. Saddleson et al. / Drug and Alcohol Dependence 149 (2015) 25–30
Table 1 Demographic characteristics among college students, by institution (N = 1437). A n (%)
B n (%)
C n (%)
D n (%)
Total n (%)
137 (9.5)
81 (5.6)
333 (23.2)
875 (60.9)
1437 (100)
23 (16.8) 22 (16.1) 92 (67.2) 20.3 (1.52)
15 (18.5) 11 (13.6) 55 (67.9) 20.7 (1.88)
162 (48.6) 85 (25.5) 86 (25.8) 18.9 (1.16)
514 (58.7) 217 (24.8) 144 (16.5) 18.7 (1.02)
718 (50.0) 338 (23.5) 381 (26.5) 19.0 (1.32)
38 (27.7) 99 (72.3)
18 (22.2) 63 (77.8)
112 (33.6) 221 (66.4)
385 (44.0) 490 (56.0)
558 (38.8) 879 (61.2)
75 (59.5) 34 (27.0) 17 (13.5)
66 (83.5) 9 (11.4) 4 (5.1)
260 (82.0) 16 (5.0) 41 (12.9)
475 (56.3) 299 (35.5) 69 (8.2)
884(64.2) 359 (26.1) 133 (9.7)
58 (42.6) 2 (1.5) 50 (36.8) 26 (19.1)
45 (55.6) 3 (3.7) 18 (22.2) 15 (18.5)
213 (64.2) 4 (1.2) 81 (24.4) 34 (10.2)
540 (61.7) 8 (0.9) 202 (23.1) 125 (14.3)
862 (60.2) 17 (1.2) 351 (24.5) 203 (14.2)
89 (65.0) 24 (17.5) 24 9 (17.5)
57 (70.4) 10 (12.3) 14 (17.3)
249 (74.8) 49 (14.7) 35 (10.5)
605 (69.2) 130 (14.9) 139 (15.9)
1007 (70.1) 215 (15.0) 214 (14.9)
N (%) Age (in years) 18 19 20–23 Mean (SD) Gender Male Female Race/ethnicity Non Hispanic White Non-Hispanic Other Hispanic Tobacco cigarette smoking status Never Former Experimenter Current E-cigarettes Never user Discontinued user Current user
p-value
Contents lists available at ScienceDirect
Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep
Risky behaviors, e-cigarette use and susceptibility of use among college students M.L. Saddleson a,∗ , L.T. Kozlowski a , G.A. Giovino a , L.W. Hawk b , J.M. Murphy c , M.G. MacLean d , M.L. Goniewicz f , G.G. Homish a , B.H. Wrotniak g , M.C. Mahoney a,e a University at Buffalo, State University of New York, School of Public Health and Health Professions, Department of Community Health and Health Behavior, Buffalo, NY, USA b University at Buffalo, State University of New York; Department of Psychology, Buffalo, NY, USA c State University of New York, at Cortland, Health Department, Cortland, NY, USA d State University of New York, Buffalo State, Department of Psychology, Buffalo, NY, USA e Roswell Park Cancer Institute, Department of Medicine and Department of Health Behavior, Buffalo, NY, USA f Roswell Park Cancer Institute, Division of Cancer Prevention and Population Sciences, Buffalo, NY, USA g D’Youville College, Center for Health Behavior Research, Buffalo, NY, USA
a r t i c l e
i n f o
Article history: Received 11 September 2014 Received in revised form 11 December 2014 Accepted 2 January 2015 Available online 24 January 2015 Keywords: Electronic cigarettes Tobacco Smoking Susceptibility to smoking Young adults College students
a b s t r a c t Background: Since 2007, there has been a rise in the use of electronic cigarettes (e-cigarettes). The present study uses cross-sectional data (2013) to examine prevalence, correlates and susceptibility to e-cigarettes among young adults. Methods: Data were collected using an Internet survey from a convenience sample of 1437, 18–23 year olds attending four colleges/universities in Upstate New York. Results were summarized using descriptive statistics; logistic regression models were analyzed to identify correlates of e-cigarette use and susceptibility to using e-cigarettes. Results: Nearly all respondents (95.5%) reported awareness of e-cigarettes; 29.9% were ever users and 14.9% were current users. Younger students, males, non-Hispanic Whites, respondents reporting average/below average school ability, ever smokers and experimenters of tobacco cigarettes, and those with lower perceptions of harm regarding e-cigarettes demonstrated higher odds of ever use or current use. Risky behaviors (i.e., tobacco, marijuana or alcohol use) were associated with using e-cigarettes. Among never e-cigarette users, individuals involved in risky behaviors or, with lower harm perceptions for e-cigarettes, were more susceptible to future e-cigarette use. Conclusions: More e-cigarette users report use of another nicotine product besides e-cigarettes as the first nicotine product used; this should be considered when examining whether e-cigarette use is related to cigarette susceptibility. Involvement in risky behaviors is related to e-cigarette use and susceptibility to e-cigarette use. Among college students, e-cigarette use is more likely to occur in those who have also used other tobacco products, marijuana, and/or alcohol. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Since 2007, electronic cigarettes (e-cigarettes) have gained popularity in the U.S. In April 2014, the Food and Drug Administration (FDA) proposed regulations restricting e-cigarette sales to minors. One public health concern is that e-cigarettes may be a “gateway,” leading e-cigarette users to begin smoking (Cobb and Abrams, 2011). E-cigarettes are often believed by users to be less
∗ Corresponding author at: 3435 Main St., Buffalo, NY 14214, USA. Tel.: +1 716 829 5702. E-mail address: [email protected] (M.L. Saddleson). http://dx.doi.org/10.1016/j.drugalcdep.2015.01.001 0376-8716/© 2015 Elsevier Ireland Ltd. All rights reserved.
harmful than cigarettes (Choi and Forster, 2013; Pearson et al., 2012). A recent review supports beliefs about reduced harmfulness, concluding “Health professionals may consider advising smokers unable or unwilling to quit through other routes to switch to e-cigarettes as a safer alternative to smoking and a possible pathway to complete cessation of nicotine use” (Hajek et al., 2014). Perceptions of e-cigarettes as less harmful than cigarettes may be associated with use of e-cigarettes. Perceived harmfulness of marijuana is strongly associated with marijuana use (Bachman et al., 1998; Bailey et al., 1992). Existing studies offer mixed results on associations between e-cigarette harm perceptions and use (Adkison et al., 2013; Choi and Forster, 2013; Sutfin et al., 2013).
26
M.L. Saddleson et al. / Drug and Alcohol Dependence 149 (2015) 25–30
Prevalence of e-cigarette ever use is highest among young adults (Adkison et al., 2013; Pearson et al., 2012; Regan et al., 2013). Young adults are more aware and have higher rates of use than older groups (Choi and Forster, 2013; Pearson et al., 2012; Pepper et al., 2013; Sutfin et al., 2013; Trumbo and Harper, 2013). Trumbo et al. surveyed students at a Colorado university in 2011: 13.0% had ever used e-cigarettes (Trumbo and Harper, 2013). Sutfin et al. surveyed undergraduates in North Carolina colleges in 2009 (in early years of e-cigarette marketing), reporting prevalence of 4.9%; those who were smokers, male, Hispanic, “other race,” or in Greek organizations (sorority/fraternity), or with lower harm perceptions of e-cigarettes were likelier to use e-cigarettes (Sutfin et al., 2013). Latimer et al. surveyed urban public universities in Texas during 2011 and reported 3.1% used e-cigarettes during the previous 30days (Latimer et al., 2013). In 2014, e-cigarette prevalence rates among middle and high school students were 6.5% (ever use) and 2.0% (current use; Dutra and Glantz, 2014). The present study measured prevalence and correlates of e-cigarette use among college students attending four colleges/universities in Upstate New York in 2013. Given concerns about e-cigarette use in never smokers of cigarettes, analyses explored correlates of e-cigarette use in never smokers. We examined factors that could lead young adults to try e-cigarettes. Among never e-cigarette users, we assessed correlates of “susceptibility” to future e-cigarette use adapting measures Pierce (Pierce et al., 1996, 2005, 1995) employed to examine ‘susceptibility to cigarette smoking’ which predicted never smokers who later become cigarette smokers (Pierce et al., 1996; Zhu et al., 2013). 2. Methods 2.1. Sample Undergraduate students in selected classes (e.g., psychology/health behaviorrelated courses) at four colleges/universities in New York State (NYS; outside of New York City) participated in fall, 2013, providing informed consent for this IRBapproved research. 2.2. Survey instrument Our 111-item, self-administered, web-based survey, used items from published studies on e-cigarettes, assessed awareness, ever and current use (i.e., within the past 30-days). Skip patterns determined items completed. We collected students’ individual school identification number and school email address, to give credit for participation and to ensure surveys were completed only once. The survey assessed demographic characteristics and other health risk behaviors, such as, cigarette use, use of other nicotine and tobacco products, patterns of alcohol use, and marijuana use. 2.3. Procedures A Web-based survey was used. At one college/university (n = 875), the survey was available through a psychology research website that was only accessible to introductory psychology students (PSY 101). At the other colleges/universities, students (n = 137, n = 81, and n = 333, respectively; 11 students did not provide university attendance information) accessed the survey website as directed by instructors. At the discretion of the investigator(s) at each college/university, respondents were granted either some form of course credit/research credits or entered into a lottery for a $25 grocery store gift card. To accurately represent a college population and because the legal age to provide consent is age 18 years, students younger than 18 years old and older than 23 were excluded from analyses (n = 111). International students were excluded from analyses (n = 100) due to cultural values which are distinct from U.S. college students and the variable availability of e-cigarettes elsewhere. 2.4. Measures 2.4.1. Demographic characteristics. Age was recoded as 18, 19 and 20–23 years to keep distributions similar, yet illustrate multiple age groups. 2.4.2. Gender. Male or female. 2.4.3. Race/ethnicity. Created from two separate items (Race and Ethnicity: “Are you Spanish, Hispanic or Latino?”), a 3 category variable was constructed: Non-Hispanic
White/Caucasian, Non-Hispanic non-whites (all races excluding white/Caucasian), and Hispanic (regardless of race). 2.4.4. School ability. Assessed by: “How well do you do in school? Would you say. . .,” “Much better than average,” “Better than average”, “Average,” “Below average” or, “Don’t know.” “Much better than average” and “better than average” were combined and then recoded into 2 categories: “Average/below average/don’t know” and “better than average”. 2.4.5. Awareness of e-cigarettes. One item asked about awareness of e-cigarettes: “Prior to today, have you ever heard about electronic cigarettes (e-cigarettes)?” Responses included: “No, I have never read anything about them and have never been told about them;” “Yes, I read a bit about them or someone told me about them;” “Yes, I am informed on the e-cig, but I have never used it;” “Yes, and I have already used an e-cigarette.” A binary variable was created (yes/no); any “yes” response indicated awareness. 2.4.6. E-cigarette ever use. Assessed by: “Have you ever tried or experimented with an e-cigarette, even one or two puffs?” Those who responded “yes” were classified as ever users. 2.4.7. E-cigarette use. Current use included use on one or more days in the previous 30 days. Respondents who ever used e-cigarettes, but not in the previous 30 days, were classified as discontinued e-cigarette users. 2.4.8. Tobacco cigarette smoking status. Based on having ever tried or experimented with tobacco cigarettes, the number of days smoked in the past 30 days, and the number of cigarettes smoked in one’s lifetime, a four category variable was created: Never smokers (never tried a tobacco cigarette), former smokers (smoked ≥ 100 cigarettes in lifetime, and have smoked 0 out of the past 30 days), experimenters (have ever tried a cigarette, have smoked < 100 cigarettes in lifetime, and have smoked 0 of the past 30 days), and current smokers (have smoked at least 1 day out of the past 30). For the multivariable analyses, smoking status was collapsed into three categories (never smokers, experimenters and ever [current and former smokers]), due to few former smokers in our sample (n = 17). 2.4.9. Other tobacco use. Assessed by one item: “From the following list, please check any of the tobacco products, besides cigarettes, you have used in the last 30 days:” (responses included: cigars, pipes, chewing tobacco, snuff, snus, hookah, clove cigarettes, bidis, other, or I have not used any other tobacco products). A binary variable was created (any/none). 2.4.10. First nicotine product used. A two-category variable was created from: Which one of the following was the first nicotine product that you used? (same responses as above 2.4.9). A three category variable was created: never used a nicotine product, first used a form of nicotine besides e-cigarettes, and first used e-cigarettes. 2.4.11. Alcohol use. Two items, (1) “During the past 30 days, how many days did you have at least one drink of any alcoholic beverage?” Responses were scored as binary (any/none), “Considering all types of alcoholic beverages, how many times during the past 30 days did you have 5 (for males)/4 (for females) or more drinks on an occasion?” Responses were recoded (any/none). 2.4.12. Marijuana use. One item, “During the last 12 months, how often did you use marijuana (cannabis, weed, pot)?” was scored as a binary variable (any/none). 2.4.13. Perceived harm of e-cigarettes. Respondents chose from a 5-category Likert scale, ranging from strongly agree to strongly disagree: “E-cigarettes are less harmful than tobacco cigarettes.” “Strongly agree” and “agree” were combined to form the “agree” group, and “strongly disagree” and “disagree” into “disagree;” a twocategory variable was then created to distinguish those who think e-cigarette use is dangerous (disagreed that e-cigarettes were less harmful than tobacco cigarettes) from those who think it is not dangerous (agreed/neutral about the harm of ecigarettes): disagree versus agree/neutral. 2.4.14. Susceptibility to tobacco cigarette smoking. This measure was adapted from Pierce’s susceptibility measure and was assessed in never users of cigarettes (Pierce et al., 1996, 2005, 1995). The measure combined two questions: “Do you think that you will try cigarettes soon?” and “If one of your best friends were to offer you a cigarette, would you try it?” Responses included: definitely yes, probably yes, probably not, definitely not and don’t know. A binary variable was created; any response other than “definitely not” qualified respondents as susceptible to cigarette smoking. 2.4.15. Susceptibility to e-cigarette use. This employed two items similar to cigarette susceptibility, except related to e-cigarettes. A binary variable was created (susceptible/not susceptible). This susceptibility measure is used to predict future e-cigarette use among those who have never tried e-cigarettes.
M.L. Saddleson et al. / Drug and Alcohol Dependence 149 (2015) 25–30 A refined e-cigarette susceptibility measure was adapted from our original measure classifying respondents as susceptible only if they, at minimum, said “probably yes” to the two susceptibility questions (excluding those responding probably not, definitely not or, don’t know); these respondents are categorized as “highly susceptible.” The distribution of ‘susceptibility to e-cigarettes’ was based on 5 categories of responses (definitely yes, probably yes, probably not, definitely not, don’t know) to two questions: (1) “Do you think you will try an e-cigarette soon?” and, (2) “If one of your best friends were to offer you an e-cigarette, would you try it?” When combined, using the traditional susceptibility definition (Pierce et al., 1995, 1996, 2005), 39.2% of never e-cigarette users are susceptible. Including only those with higher susceptibility (responded definitely yes and/or probably yes), our alternate measure estimated 10.1% of never e-cigarette users as highly susceptible. 2.5. Statistical analysis Descriptive statistics were used to summarize results (Stata version 13, Stata Corporation, College Park, TX). Chi-square tests assessed differences between institutions, e-cigarette ever versus never users, and e-cigarette current versus not current users. Multivariate logistic regression models evaluated the independent influence of demographic variables, selected health behavior and belief variables on ever use of e-cigarettes and concurrent use of e-cigarettes smoked cigarettes. These same independent variables were used to predict susceptibility to both use of tobacco cigarettes and e-cigarettes. Adjusted odds ratios (AOR) and 95% confidence intervals (CI) were computed; two-sided alpha level was 0.05. Multivariable models controlled for demographic variables (age, gender, and race/ethnicity) and other variables represented in the models.
3. Results Respondents included 1437 students (18–23 years) from four colleges/universities. See Table 1. There are differences between colleges/universities for age, gender, race/ethnicity (p < 0.001) and cigarette smoking status (p = 0.001). Overall, 95.5% of respondents report awareness of e-cigarettes; 29.9% are ever users and 14.9% are current users of e-cigarettes, with 6.4% reporting concurrent use of both e-cigarettes and tobacco cigarettes. 3.1. Correlates of ever use of e-cigarettes Table 2 presents descriptive statistics and multivariable analyses for both ever use and current use of e-cigarettes. Lower odds of ever use of e-cigarettes are seen among persons 20–23 years old, females, non-Hispanic non-whites, and those reporting better than average school performance. Those who have experimented with cigarette smoking have a 5-fold greater odds of having ever used e-cigarettes (AOR = 5.52, 95% CI = 3.89–7.85) and ever smokers of cigarettes have nearly 20 times greater odds of having ever used e-cigarettes (AOR = 19.10, 95% CI = 12.12–30.12). Ever use of ecigarettes is also higher among those who used any other tobacco products beside cigarettes in the past 30 days (AOR = 2.38, 95% CI = 1.72–3.29), those who used marijuana in the past 12 months (AOR = 2.37, 95% CI = 1.69–3.31) and those with lower e-cigarette harm perceptions (AOR = 1.78, 95% CI = 1.10–2.86). 3.2. Correlates of current use of e-cigarettes Lower odds of current e-cigarette use are seen in persons 20–23 years old, females, non-Hispanic non-whites, and those reporting better than average school performance. Those who experimented with cigarettes (AOR = 3.12, 95% CI = 1.99–4.91) or those who were ever cigarette smokers (former and current smokers) (AOR = 6.60, 95% CI = 4.09–10.63) have greater odds of current e-cigarette. Current e-cigarette use is also more likely among students who reported any binge drinking in the past 30 days (AOR = 1.71, 95% CI = 1.01–2.90), used tobacco products other than cigarettes during the past 30 days (AOR = 3.11, 95% CI = 2.13–4.54), and have lower perceptions about the harms of e-cigarettes (AOR = 1.34, 95% CI = 1.01–2.90).
27
To explore correlates of e-cigarette use among persons with relatively little cigarette experience, multivariable models assessed e-cigarette ever use and current use among a sub-sample of those who have smoked less than 100 tobacco cigarettes in their lifetime (n = 1257), which included never smokers (64.7%), experimenters (26.8%), and current smokers (8.5%) (those reporting smoking < 100 cigarettes ever, but had smoked within the past 30 days). Results among this subsample are similar to those for the entire sample, suggesting that our findings primarily apply to those who are not yet established tobacco cigarette smokers. However, one notable difference among this subgroup (for both ever and current users of e-cigarettes), is a significant effect for marijuana use (ever use AOR = 3.15, 95% CI = 2.30–4.33; current use AOR = 1.90; 95% CI = 1.26–2.85). 3.3. Susceptibility to tobacco cigarette smoking A logistic regression model was examined among never cigarette smokers (n = 862) to determine whether e-cigarette use was correlated with susceptibility to initiating smoking of tobacco cigarettes, controlling for demographics and first nicotine product used. Among never smokers (n = 862), in terms of first nicotine product used, 70.0% had never used a nicotine product, 27.7% had used some other form besides e-cigarettes first. The highest reported usage was hookah (14.4%), cigars followed (9.7%), and 2.3% reported using an e-cigarette first (other products had rates ≤2.0%). The majority of ever smokers used some other form of nicotine besides e-cigarettes as their first product, with only 2.5% having used e-cigarettes first. Our findings indicate that e-cigarette use is not related to susceptibility to smoking cigarettes, rather, e-cigarette users more commonly first used another nicotine product (86.9%) compared with first using e-cigarettes (7.2%) (p < 0.001) [5.9% had never used any nicotine product]. 3.4. Susceptibility to e-cigarette use Table 3 examines susceptibility to e-cigarette use among never users of e-cigarettes. Never e-cigarette users who were younger (18 or 19 years old) are more likely to be susceptible to e-cigarette use, as are those who drank any amount of alcohol in the past 30 days or, reported any past-month binge drinking. Those who used other tobacco products besides cigarettes in the past 30 days, those who had experimented with smoking tobacco cigarettes or, were ever tobacco cigarette smokers, and those who had lower perceptions about the harms of e-cigarettes were each more likely to be susceptible to e-cigarette use. Using the alternate susceptibility measure to evaluate higher levels of susceptibility to e-cigarette use, never e-cigarette users who identified themselves as non-Hispanic other race (AOR = 2.20, 95% CI = 1.23–3.95), those who used marijuana in the past 12 months (AOR = 2.63, 95% CI = 1.49–4.62), were ever cigarette smokers (AOR = 15.44, 95% CI = 6.91–34.50), had ever experimented with cigarettes (AOR = 3.12, 95% CI = 1.79–5.45), or current users of other tobacco products (AOR = 1.74, 95% CI = 1.01–2.99) have greater odds of being susceptible to using e-cigarettes. As found previously, males and current or former cigarette smokers (18–23 years old) are more likely to ever use e-cigarettes (Adkison et al., 2013; Choi and Forster, 2013; Goniewicz and Zielinska-Danch, 2012; Regan et al., 2013). Results on race/ethnicity are mixed (Choi and Forster, 2013; Pearson et al., 2012; Regan et al., 2013; Sutfin et al., 2013). Sutfin et al. (2013) found Hispanics and those of “other” races, compared to Whites, to be about twice as likely to have used e-cigarettes; while we found those of nonHispanic “other” (non-White) races are one-half as likely to have ever used e-cigarettes compared to Whites. These differences could relate to yet unidentified geographic differences.
28
M.L. Saddleson et al. / Drug and Alcohol Dependence 149 (2015) 25–30
Table 1 Demographic characteristics among college students, by institution (N = 1437). A n (%)
B n (%)
C n (%)
D n (%)
Total n (%)
137 (9.5)
81 (5.6)
333 (23.2)
875 (60.9)
1437 (100)
23 (16.8) 22 (16.1) 92 (67.2) 20.3 (1.52)
15 (18.5) 11 (13.6) 55 (67.9) 20.7 (1.88)
162 (48.6) 85 (25.5) 86 (25.8) 18.9 (1.16)
514 (58.7) 217 (24.8) 144 (16.5) 18.7 (1.02)
718 (50.0) 338 (23.5) 381 (26.5) 19.0 (1.32)
38 (27.7) 99 (72.3)
18 (22.2) 63 (77.8)
112 (33.6) 221 (66.4)
385 (44.0) 490 (56.0)
558 (38.8) 879 (61.2)
75 (59.5) 34 (27.0) 17 (13.5)
66 (83.5) 9 (11.4) 4 (5.1)
260 (82.0) 16 (5.0) 41 (12.9)
475 (56.3) 299 (35.5) 69 (8.2)
884(64.2) 359 (26.1) 133 (9.7)
58 (42.6) 2 (1.5) 50 (36.8) 26 (19.1)
45 (55.6) 3 (3.7) 18 (22.2) 15 (18.5)
213 (64.2) 4 (1.2) 81 (24.4) 34 (10.2)
540 (61.7) 8 (0.9) 202 (23.1) 125 (14.3)
862 (60.2) 17 (1.2) 351 (24.5) 203 (14.2)
89 (65.0) 24 (17.5) 24 9 (17.5)
57 (70.4) 10 (12.3) 14 (17.3)
249 (74.8) 49 (14.7) 35 (10.5)
605 (69.2) 130 (14.9) 139 (15.9)
1007 (70.1) 215 (15.0) 214 (14.9)
N (%) Age (in years) 18 19 20–23 Mean (SD) Gender Male Female Race/ethnicity Non Hispanic White Non-Hispanic Other Hispanic Tobacco cigarette smoking status Never Former Experimenter Current E-cigarettes Never user Discontinued user Current user
p-value
