Study sample and protocols

The Vaping and Patterns of E-cigarette Use Research (VAPER) study is a web-based longitudinal cohort study of adults (aged ≥21 years) in the US who used ENDS and ENNDS (Electronic Non-Nicotine Delivery Systems) at least five days/week¹⁶. Nearly all respondents (98%) reported using ENDS (containing nicotine). This study used data from waves 2 (December 2020–April 2021) and 5 (February–April 2023) of the VAPER study. Recruitment utilized a Craigslist-focused strategy, posting ads on gigs and job boards. Eligible participants provided verifiable identifying data, including name, date of birth, mailing address, and cell phone number; they also completed online surveys on REDCap, reporting their ENDS use patterns and behaviors. They also submitted photos of their most frequently used ENDS device and liquid. Survey responses were utilized when device and liquid characteristics were unavailable from photos. Rigorous data review and cleaning procedures were implemented after coding to ensure the data quality and authenticity of submitted photos. Detailed information about the protocols has been reported¹⁶. Follow-up survey invitations were sent to people’s mobile phones and email addresses, and the links to the survey were tied to their previously established record ID number if they validly responded and expressed interest in participating in subsequent waves. To maintain eligibility in follow-up waves, participants were required to report continued use of ENDS for at least five days per week.

The survey question regarding the primary reason for continuing to use ENDS was modified prior to wave 2. To examine the longitudinal change in primary reason for continued usage, we used data in wave 2 (December 2020–April 2021) as the 1st timepoint and wave 5 (February–April 2023) data as the 2nd timepoint in this study.

In wave 2, 1187 participants validly responded to the survey. Among these, 526 indicated an interest in staying in the study, reported still using ENDS ≥5 days per week, and provided valid responses in wave 5; this is the study sample. There was no significant difference between the study sample and other participants in wave 2 in terms of sociodemographic characteristics (including age, gender, race, income, region, and Hispanic origin) (p>0.05). The Virginia Commonwealth University Institutional Review Board (IRB) approved the study (HM20015004), with the Johns Hopkins Bloomberg School of Public Health IRB relying on the IRB of VCU as the IRB of record (IRB9277).

Measurements

Primary reason for continuing to use ENDS

Participants were asked to select one reason as the primary reason for continuing to use ENDS from options: ‘To quit smoking’, ‘To cut down smoking’, ‘To use when I’m not allowed to smoke’, ‘Absence of smell’, ‘Because I enjoy the flavor’, ‘Because I enjoy the boost’, ‘Curiosity/Just want to try them’, ‘Because I am addicted’, and ‘Some other reason (including stress, anxiety, depression, or other mental health concerns, etc.)’. We classified those options into four categories: 1) quitting/reducing smoking (‘To quit smoking’, ‘To cut down smoking’); 2) addiction (‘Because I am addicted’); 3) flavor (‘Because I enjoy the flavor’); and 4) other reasons (‘To use when I’m not allowed to smoke’, ‘Absence of smell’, ‘Because I enjoy the boost’, ‘Curiosity/Just want to try them’, ‘Because of stress, anxiety, depression, or other mental health concerns’, and ‘Some other reason’), then created a dichotomous variable (1=Yes, 0=No) for each category. A sensitivity test showed that there was no significant difference between combining ‘enjoy the boost’ with ‘addiction’ and combining it with ‘other reasons’.

Device-liquid grouping

We classified ENDS by examining the combination of the device (reusability, refillability, uses a tank or pod, modifiability of the power or airflow or coil) and liquid (freebase and nicotine salt) characteristics that can affect user nicotine and toxicant exposure^3,17. A device was classified as having adjustable settings if it allowed users to modify the power, coil, or airflow. We coded device and liquid information directly from submitted photos and searched the internet to identify the characteristics of the device (e.g. brand, model) and liquid (e.g. nicotine formulation and concentration). The self-report data were used when device or liquid characteristics were not available from submitted photos or internet searches. Over 90% of participants reported using one of five device-liquid groupings in waves 2 and 5: 1) disposable device (salt, no adjustable settings); 2) disposable pod (salt, no adjustable settings); 3) refillable pod (freebase, adjustable settings); 4) refillable pod (salt, adjustable settings); and 5) tank (freebase, adjustable settings).

Other variables

Participants self-reported their geographical location (Northeast, Midwest, South, West), gender (man, woman, non-man/woman, prefer not to answer), age (21–29, 30–44, ≥45 years), age of first use of ENDS (<21, ≥21 years), race (White, other single race, multi-racial, prefer not to answer), ethnicity (Hispanic/Latino, non-Hispanic/Latino, prefer not to answer), income ($) (0–39999, 40000–59999, 60000 –99999, ≥100000, prefer not to answer), and sexual identity (heterosexual or straight, non-heterosexual or non-straight, prefer not to answer). Smoking status was based on two items: 1) ‘How many cigarettes have you smoked in your entire life?’; and 2) ‘Have you smoked a cigarette in the past 30 days?’. Participants were classified as people who have never smoked if they smoked <100 cigarettes in their lifetime; as people who formerly smoked if they smoked ≥100 cigarettes in their lifetime and did not smoke in the past 30 days; and as people who currently smoke if they smoked ≥100 cigarettes in their lifetime and smoked in the past 30 days. ENDS dependence was measured by a 4-item E-cigarette Dependence Scale (EDS) score (range: 0–4, 4 being most dependent)¹⁸.

Statistical analysis

All analyses were conducted using SAS (Version 9.4, SAS Institute, Cary, NC). Descriptive statistics were provided for sociodemographic variables, smoking status, and ENDS device-liquid grouping for the overall sample and by reasons for continued ENDS usage, in each wave. The Rao-Scott chi-squared test was used to assess differences in categorical variables, and the McNemar test was used to determine if the proportions of categories in the two waves differed significantly. Associations between primary reason for continued ENDS usage (quitting/reducing smoking, addiction, flavor, and other reasons) and device-liquid grouping, were determined through multivariable logistic regression models. Adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were calculated and reported for independent variables. These models controlled for participants’ sociodemographic characteristics (age, race, and sexual orientation) and smoking status.

Based on data from the 2019 Tobacco Use Supplement to the Current Population Survey (TUS-CPS), post-stratification survey weighting was employed for gender, age, and race, with weights ranging from 0.33 to 1.80. This ensured the representativeness of US adults who frequently use ENDS. Weighted frequencies (n), scaled to our sample size, are reported for all analyses. P-values were adjusted to accommodate the large sample size resulting from applying survey weights. The pairwise deletion method handled missing data (<2% of participants). A 2-sided p<0.05 was considered statistical significance.

Factors associated with primary reason for continued ENDS usage in waves 2 and 5

The primary reason for continued ENDS usage was significantly associated with age, age of first ENDS use, smoking status, and device-liquid grouping. In wave 2, participants who were younger (<30 years vs ≥45 years) or first used ENDS aged <21 years were less likely to report quitting or cutting down smoking as their primary reason for use (31.0% vs 53.8%, p=0.002; 32.5% vs 46.5%, p=0.022, respectively); and more likely to report addiction as the reason (41.4% vs 24.4%, p=0.003; 41.8% vs 26.1%, p=0.004, respectively) compared to their counterparts (Table 1). The age difference in the primary reason for continued ENDS usage may relate to their smoking status. Nearly all participants aged ≥45 years either currently or formerly smoked, which was significantly higher compared to participants aged <30 years (98.4% vs 75.0%, p<0.0001). Participants who have never smoked were more likely to report addiction as their primary reason for continued usage compared to those who formerly smoked or currently smoke (48.0% vs 28.2% vs 21.2%, respectively, p=0.003) (Table 2). Participants reporting addiction as the primary reason had significantly higher ENDS dependence compared to participants reporting other reasons (2.8 vs 2.1, p<0.0001). These significant associations were also found in wave 5 (Tables 1 and 2).

Associations between primary reasons for continued ENDS usage and device-liquid grouping

A set of multivariable logistic regression models was utilized to examine the associations between the four primary reasons for continued ENDS usage (quitting/reducing smoking, addiction, flavor, other reasons) and device-liquid grouping in each wave after controlling for age, race, sexual orientation, and smoking status (Table 3). Compared to those using tanks (freebase, adjustable settings), participants using disposable devices (salt, no adjustable settings) were less likely to report quitting or reducing smoking as the primary reason (W2: AOR=0.34; 95% CI: 0.14–0.80; W5: AOR=0.52; 95% CI: 0.28–1.00); and participants using device-liquid groupings with a nicotine salt liquid were more likely to report addiction as the primary reason (disposable device [salt, no adjustable settings]: W2: AOR=2.91; 95% CI: 1.33–6.36, W5: AOR=3.50; 95% CI: 1.82–6.73; disposable pod [salt, no adjustable settings]: W2: AOR=2.15; 95% CI: 1.12–4.15, W5: AOR=3.89; 95% CI: 1.85–8.19; refillable pod [salt, adjustable settings]: W2: AOR=2.47; 95% CI: 1.26–4.82, W2: AOR=2.67; 95% CI: 1.25–5.70). Regarding flavor as the primary reason for continued usage, participants using disposable devices (salt, no adjustable settings) were less likely to report flavor as the primary reason in wave 2 (AOR=0.12; 95% CI: 0.03–0.44).

From wave 2 to 5, there was a significant change in the primary reason for continued ENDS usage, with fewer participants reporting quitting or cutting down smoking (43.5% vs 32.1%, p<0.001) and more reporting addiction (29.2% vs 34.6%, p=0.014) (Figure 1). The percentage of people reporting flavor as the primary reason remained nearly the same (12.8% vs 11.1%, p=0.292). An increase in participants using disposable devices (salt, no adjustable settings) (7.9% vs 25.2%) was also observed. At the same time, there was a decrease in using tanks (freebase, adjustable settings) (37.5% vs 27.4%), disposable pods (salt, no adjustable settings) (18.6% vs 13.2%), refillable pods (salt, adjustable settings) (18.4% vs 13.0%), and refillable pods (freebase, adjustable settings) (13.8% vs 11.8%) (Table 2). The association between the change in the primary reason for continued ENDS usage and device-liquid grouping was explored. Among participants who reported quitting/reducing smoking or addiction as the primary reason in wave 2, more than half reported the same reason in wave 5 (55.7% and 70.4%, respectively). Compared to those who changed from addiction to other reasons, participants who consistently reported addiction as their primary reason for continued ENDS usage were less likely to use tanks (freebase, adjustable settings) (14.2% vs 41.9%) and more likely to use disposable pods (salt, no adjustable settings) in wave 5 (18.9% vs 6.2%) (p=0.010) (Table 4). Compared to those changing from flavor to other reasons, a greater proportion of participants who consistently reported flavor as the reason, used tanks (freebase, adjustable settings) in wave 5 (50.6% vs 31.1%) and a smaller proportion used disposable devices (salt, no adjustable settings) (6.2% vs 19.8%) or refillable pods (salt, adjustable settings) (5.9% vs 26.2%). However, the difference was not statistically significant (p=0.087).

Figure 1

Change in primary continued ENDS usage reasons from wave 2 to 5, VAPER cohort study, wave 2 (December 2020–April 2021) and wave 5 (February–April 2023) (N=526)

Limitations

This study should be interpreted within the context of its limitations. First, participants in this study used ENDS ≥5 days per week in both waves. Thus, findings cannot be generalized to a population with less frequent ENDS usage or a population using ENNDS. The sample could be biased if individuals who successfully quit smoking using ENDS then reduced or stopped using them (<5 days/week). Second, to maintain a manageable survey length for participants, detailed device and liquid data were collected and reported only for participants’ most commonly used device and liquid, not accounting for other devices and liquids they might use. Third, we grouped ‘To use when I’m not allowed to smoke’, ‘Absence of smell’, ‘Because I enjoy the boost’, ‘Curiosity/Just want to try them’, and ‘Some other reason (including stress, anxiety, depression, or other mental health concerns, etc.)’ together as ‘Other reasons’ for continued ENDS usage due to the small sample size of these options and analysis purposes. Results of ‘Other reasons’ provide limited information to inform regulatory efforts. And participants self-reported the primary reason for continued ENDS usage. There could be some reporting bias. In addition, the retention rate for wave 2 to wave 5 is 44.3%. Although there were no significant differences in sociodemographic characteristics between the study sample and other participants in wave 2 who did not respond to wave 5 survey, there could be a potential attrition bias.