Sample and procedure

Data was obtained between May and August 2014 by using a Dutch online survey. Inclusion criteria were being a smoker and/or vaper, being between 18 and 65 years old, and being aware of the existence of e-cigarettes. Respondents were recruited via internet by Facebook advertisements and calls on (e-)cigarette related discussion boards. In addition, 200 flyers were distributed among different public places.

The online questionnaire was based on previous research using the I-Change Model and a previous German study⁴⁰. A total of 970 people visited the website. From this group, 49 respondents did not meet the inclusion criteria, 84 respondents did not answer the informed consent question, and 69 respondents answered the informed consent question negatively. This resulted in 768 respondents who started the questionnaire (79% of the visitors), and a total of 477 complete cases (49 % of the visitors and 62% of the people who actually started the questionnaire). The questionnaire took approximately 30 minutes to complete. Among respondents who completed the questionnaire 20 gift certificates to the value of 25 euro were raffled.

The study was approved by the Medical Ethics Committee of the Atrim-Orbis-Zuyd Hospital (14-N-63).

Theoretical framework

The Integrated Model for Change (I-Change model) was used in order to assess relationships with perceived pros and cons of vaping, social modelling and social norms towards vaping, and self-efficacy towards vaping. The I-Change Model assumes three phases in the behavioural change process: awareness (determined by cognizance, knowledge, risk perceptions, and cues to action), motivation (determined by attitudes, social influence beliefs and self-efficacy) and action (determined by self-efficacy, intentions, action planning and plan enactment). Additionally, these phases are influenced by distal preceding factors (such as personality, social environmental factors, and information factors (such as the type and quality of the health communication provided)^41–44. This study focused on the motivational determinants as specified by the I-Changed Model.

Demographics

Demographic information was assessed by seven items, pertaining gender, age, nationality, marital status (1=single; 2=in a relationship/cohabiting; 3=married; 4=divorced; 5=widow(er); 6=other), and the highest completed level of education (1=low [no education, primary school, lower vocational education and general secondary education]; 2=middle [secondary vocational education and secondary education]; 3=high [higher education and university education]). The total amount of monthly income (after the subtraction of taxes) was assessed by three categories (1=low [up to €1300]; 2=middle [€ 1300 to 2400]; 3=high [€ 2400 and above]). Employment status was assessed by two categories (1=Employed [entrepreneur, wage labour, and employment for government]; 2=Unemployed [incapacitated, unemployed, retired, student, and homemaker]).

General smoking/vaping behaviour

Respondents’ general smoking/vaping behaviour was assessed by two items; ‘I smoke/vape..’ (1=both conventional cigarettes and e-cigarettes; 2=solely conventional cigarettes; 3=solely e-cigarettes) and ‘I..’ (1=first smoked conventional and later began to vape e-cigarettes; 2=first vaped e-cigarettes and later began to smoke conventional cigarettes; 3=smoke conventional cigarettes and never vaped e-cigarettes; 4=vape e-cigarettes and never smoked conventional cigarettes; 5=first smoked conventional cigarettes, and have vaped e-cigarettes in-between; 6=first vaped e-cigarettes, and have smoked conventional cigarettes in-between).

Smoking behaviour

Smokers and dual users received three questions regarding their smoking behaviour. These questions assessed the frequency of smoking cigarettes (‘How often do you smoke?’), onset (‘Since when do you smoke?’) and number of cigarettes smoked (‘How many cigarettes do you smoke on average per day?’).

Vaping behaviour

Vapers and dual users received five questions regarding their vaping behaviour. These questions assessed the frequency of vaping e-cigarettes (‘How often do you vape?’), onset (‘Since when do you vape?’), reasons to start (‘What was your main reason to start to use the e-cigarette?’), how respondents became familiar with e-cigarettes (‘How did you become familiar with the e-cigarette?’), and used cartridges (‘I vape..’(nicotine-free versus nicotine-containing cartridges)).

Current quit attempt

Two items assessed quitting behaviour of conventional cigarettes. First, respondents were asked if they were currently attempting to quit (not smoking for at least 24 hours) conventional cigarettes (1=no; 2=yes). Secondly, respondents who were making a quit attempt were asked since when they were making this attempt (open question: number of years/months/weeks/days).

Attitudes towards e-cigarettes

Attitudes towards e-cigarettes were assessed by 25 items on a five point Likert scale about perceived advantages and 23 items on perceived disadvantages of e-cigarettes. An overview of the items is depicted in Table 2 (advantages) and Table 3 (disadvantages).

Due to small variances in both constructs, items were dichotomised. Perceived advantages (α = .93) were recoded into either 0=totally disagree/disagree/neutral or 1 =agree/totally agree. Perceived disadvantages (α = .92) were recoded into either 0=totally agree/agree/neutral or 1=disagree/totally disagree.

Social Influences

Social influence was measured by eight social modelling questions and eight social norm questions. First, respondents were asked to indicate whether their partner smoked conventional cigarettes (0=no/non-applicable; 1=yes), and how many of their family members, friends, and colleagues smoked conventional cigarettes (1=all; 2=more than half; 3=half; 4=less than half; 5=none; 6=non-applicable). The same four questions were asked regarding vaping e-cigarettes.

Due to small variances in both constructs, items were dichotomised into 0 (receiving no modelling, which included ‘non-applicable’) or 1 (receiving modelling). Scales were calculated by sum scores (modelling conventional cigarettes: α = .45, modelling e-cigarettes: α = .56).

Social norm was also assessed by two constructs. ‘Would your (partner/family member/friends/ colleagues) prefer that you vape e-cigarettes instead of conventional cigarettes (α = .95)’ and ‘Does your (partner/ family member/friends/ colleagues) think vaping e-cigarettes is good? (α = .68)’. Questions were answered on a six point scale (1=totally disagree; 5=totally agree; 6=non-applicable).

Self-efficacy

Self-efficacy beliefs about not vaping in potential difficult situations were assessed by 16 items (α = .95) on a five point Likert scale (see Table 4 for an overview). Self-efficacy about not to vape was measured among vapers, dual users and smokers. When assessing self-efficacy, it is custom to ask both participants engaging and not-engaging is a particular behaviour to provide their self-efficacy ratings.

Intention to quit electronic and conventional cigarettes

The intention to quit electronic and/or conventional cigarettes was assessed by two items. Respondents indicated on a scale from 1 (very weak) to 10 (very strong), how strong they were intending to stop 1) conventional cigarettes and 2) e-cigarettes.

Data analysis

The three groups (smokers, vapers and dual users) were classified based on how respondents considered themselves according to the first general smoking/vaping behaviour question. General descriptive analyses were performed to describe the whole sample and the three groups. Group differences regarding sample and vape characteristics were identified by performing Chi-Square tests. ANOVA analyses assessed differences between vapers and dual users (see table 1 for included independent variables). Differences between smokers, vapers and dual users were tested by post hoc pairwise comparisons by Tukey HSD tests (see table 2, 3 and 4 for the included independent variables).

Multinomial logistic regression analyses were performed to identify associations between socio-demographic variables and the constructs from the I-Change model with respondents’ smoking behaviour. In the regression analyses smoking behaviour (either: smoker, vaper or dual user) was used as dependent variable and only included complete cases (see table 5 for the included independent variables). Items that were answered with ‘I don’t know’ were recoded as missing value. In total, 60 of the 477 respondents were excluded from the analyses, because they had a missing value on any of the variables. This resulted in 104 smokers, 240 vapers and 73 dual users included in the regression analyses.

All analyses were conducted with IBM SPSS statistics (version 20).

Sample description

Table 1 provides an overview of the sample characteristics. In total 259 (54%) vapers, 135 (28%) smokers and 83 (18%) dual users completed the questionnaire. The questionnaire was completed by 185 men, and respondents had an average age of 37.1 years. Most (95%) of the respondents were Dutch, 3.6% had the Belgian nationality and 1.4% had a different nationality. Vapers more often had a higher income and paid work in comparison to smokers.

Only 5 of the 259 vapers (1.9%) had never smoked conventional cigarettes. The other 254 vapers indicated that they had smoked conventional cigarettes in the past, but successfully quit smoking. Almost 23% of the dual users and 10.4% of the smokers indicated to be currently involved in a quit attempt. Yet, additional analyses concerning the quit attempt for both dual users and smokers revealed that these quit attempts were unsuccessful.

Table 1 also shows that vapers were significantly more often daily users and had used e-cigarettes for a longer period of time than dual users. Although group differences were found, the majority among both groups indicated to use mainly or solely nicotine-containing cartridges.

Significant differences in reasons to start e-cigarette use were found as well. Among 60% of the vapers and 27% of the dual users, the most frequently mentioned most important reason for use was believing that the product would be better for their health. Using e-cigarettes as smoking cessation tool was indicated to be the second most important reason for use. Dual users indicated ‘to limit the amount of conventional cigarettes’ as a third important reason for use.

Motivational factors Perceived advantages

Table 2 shows that there were significant group differences regarding mean scores of agreement on 20 of the 25 statements about advantages of e-cigarettes. Smokers scored lowest on the 20 items and were least convinced of the product, while vapers scored the highest and were most convinced. Smokers perceived vaping, in comparison to both groups, significantly less often to be cheaper than smoking conventional cigarettes. Vapers perceived vaping significantly more often to be just as tasty as smoking, in comparison to both other groups.

Perceived disadvantages

Table 3 shows that there were significant group differences regarding mean scores of agreement on 14 of the 23 statements about disadvantages of e-cigarettes. The highest scores on the disadvantages scale were found among the smokers while vapers reported the lowest scores.

Smokers perceived vaping significantly more often as costly, time-consuming, unhandy, causing nausea, dizziness, and/or mouth and throat irritations. Vapers perceived vaping, compared to both other groups, significantly less often to be shifting the nicotine-dependence and leading to increases in toxic substances in their body. Vapers perceived vaping, in comparison to smokers, significantly less often to be addictive.

Social influences

An overview of the social influence items and the differences between the three groups can be found in Table 4.

Significant differences between the groups were found on all eight items of the two social norm constructs. Vapers scored the highest, indicating that their social environment had a more positive attitude towards e-cigarettes compared to the smokers who scored the lowest.

Furthermore, vapers experienced significantly less social modelling for the conventional cigarette from their partner, family members, and friends compared to dual users and smokers. Also, the vapers experienced less modelling for the conventional cigarette from their colleagues compared to the smoker group. In all four items regarding social modelling for e-cigarettes, smokers scored lower in comparison to the dual and vapers.

Self-efficacy

Table 4 illustrates significant group differences on four of the 16 self-efficacy items. Vapers indicated to have the highest self-efficacy, whereas smokers scored the lowest.

Vapers indicated to be significantly more confident to manage not to vape when they see someone vaping, when they are sad, stressed, tensed, when they crave for nicotine, or wake up in the morning, compared to smokers. Vapers indicated to be significantly more confident not to vape when an e-cigarette is offered, when they miss the flavour of conventional cigarettes, and when they just ate.

Smokers indicated to be significantly less confident not to vape when they do not know how to act, when they are bored, and when they had a break from work.

Intention to quit conventional cigarettes. Smokers had a significantly higher intention to quit conventional cigarettes (M=6.2) than dual users (M=4.9; (F=16.2; p < 0.001).

Factors associated with smoking/vaping behaviour

Table 5 presents the results of the multinomial regression analyses (N.B. results can differ from our ANOVA analysis due to possible suppressor effects, i.e. the regression analysis corrected the correlations with other included variables). Results revealed that young respondents were more likely to smoke or to smoke in combination with vaping instead of solely vaping e-cigarettes. Older respondents were on the other hand more likely to be dual users instead of smokers. Furthermore, respondents with a higher level of education or higher income levels were more likely to smoke cigarettes than solely vape e-cigarettes. Respondents with low levels of education were more likely to be dual users instead of smokers.

Respondents who perceived fewer advantages of e-cigarettes were more likely to be smokers or dual users than vapers. Respondents who indicated to perceive more advantages were more likely to be dual users than smokers. Similar results were found for the perceived disadvantages: respondents who indicated to perceive more disadvantages were more likely to be smokers than vapers. People who perceived fewer disadvantages were more likely to be dual users than smokers. People with lower self-efficacy values were more likely to be either smokers or dual users.

Finally, people who received more social modelling for conventional cigarettes were more likely to be dual users instead of vapers. People who received less social modelling for e-cigarettes were more likely to be smokers than vapers. People who perceived more social modelling were more likely to be dual users than smokers.

Strengths and limitations

The explorative, theory-driven nature of the study can be considered as one of it its important strengths. The distinction between vapers and dual users is not often made. Further, the study examined a broad range of motivational determinants associated with e-cigarette use. These insights offer the opportunity to optimize health communication about e-cigarettes by providing tailored messages to the opinions of the various groups of (potential) users.

This study however also is subject to some limitations. First, most vapers were recruited through online forums. This may have led to an overrepresentation of dedicated users, who may be more positive towards e-cigarette use than non-responders. Also, the provided incentives to respondents could have biased the research sample. Second, as the goal of this study was to study motivational differences between the study groups and not to study e-cigarette prevalence, no conclusions about e-cigarette use on a population level can be drawn from this study. Third, the sample size, especially among dual users and smokers was modest. The sample size was also affected by a non-completion rate of almost 40%, probably caused by the length of the questionnaire. Fourth, we only focused on motivational factors influencing e-cigarette use. Potential other factors of influence (e.g. advertisements of the product or influences of the media or other external factors) are not examined is this study. Fifth, our results are cross-sectional, so no inferences about factors predicting e-cigarette use can be made. Finally, we did not assess pre-motivational and post-motivational factors. Yet, assessing factors that determine awareness of e-cigarette use are also needed as well as studies targeting post-motivational factors such as action planning (preparatory and coping plans) and plan enactment. For instance, a recent study on smoking cessation demonstrated that smokers‘ intentions to quit were to a large extent explained by action plans and plan enactment⁴².