Predictors of Persistent Waterpipe Smoking Among University Students in The United States

Smoking continues to be a significant public health problem causing 443,000 premature deaths annually in the United States [1] and 5.4 million worldwide [2]. It is currently the leading preventable cause of premature death both in the United States (US) and worldwide [2,3], and is projected to remain as such in 2020 [4]. Smoking annually costs the US society approximately $193 billion with $96 billion in direct medical costs [1].

Waterpipe smoking is a recent form of tobacco use in the US and is growing in popularity especially among the adolescents and young adults [5] with a misconception of relative safety [5-8] and variety of sweetened flavors [5,9]. Waterpipe smoking is common in the Middle East, India, and segments of Asia, and increasingly becoming more prevalent in Western countries [10]. Waterpipes have been referred to as the following names: Shishah [5,9,10], hookah [5,9,10], argeela, Goza [5,10], hubble-bubble [5,10], and narghile [9,10].

The American Lung Association (ALA) [11], the World Health Organization (WHO) [12], and the American Cancer Society (ACS) [6] have warned against health risks associated with waterpipe smoking. Similar to cigarettes, it can increase the risk of cancer [5,9,13- 15], cardiovascular disease [5,9,13], pulmonary disease [5,9,14-16], and poor fetal outcomes [5,14,15]. Furthermore, using the same mouthpieces in social groups can also spread communicable diseases [5,14].

Previous studies have also demonstrated that smoking tobacco using a waterpipe causes exposure to carbon monoxide (CO), smoke [6,17] nicotine [17], as well as carcinogens including hydrocarbon and heavy metals [6,18,19] and tar [14] and can result in nicotine dependence [5,9,13,14].

Because of the rising prevalence of hoohkah bars in close proximity to college campuses in the last 10 years [5,8,20,21], prevention and intervention strategies that combat the spreading of this health hazard are greatly needed. Although a number of studies have examined the predictors of a past ever use or past month use of a waterpipe to smoke tobacco among US college students [9,22,23], predictors of a continued persistent use have yet to be explored. Many students in colleges may have tried a waterpipe in the past month or in their lifetime but did not become habitual waterpipe smokers. Since possible addiction and associated health risks are expected in habitual waterpipe smokers, our objectives were to identify predictors of persistent waterpipe smoking among college students who have attempted to use a waterpipe to smoke tobacco. Knowledge of these predictors is critical for designing effective prevention strategies to prevent further diffusion of waterpipe smoking into the society and to establish interventions that can help existing waterpipe smokers quit.

A cross-sectional study was conducted using data collected from University of Houston (UH) students via an online survey constructed with Qualtrics. In February/2011, an email was sent to the UH student body inviting them to participate in the research. The email directed students to an online link which included an informed consent explaining that their participation is voluntary and anonymous. After clicking on the appropriate box to accept, the survey became accessible. To increase the response rate, all participants were placed in a drawing for a chance to win one of ten $50 Starbucks gift cards. After completion of the survey, the participants were instructed to click on another link that opened a separate page to enter their email addresses. This allowed the participants to record their email for the purpose of the random drawing but maintained the anonymity of the unlinked survey data. The protocol was approved by the Institutional Review Board (IRB) at the University of Houston. Data collection was completed by the end of March, 2011.

The survey was adapted from surveys used in previous studies [22- 24] with questions on demographic characteristics (including gender, age, race/ethnicity), tobacco use, perception of risk, perception of addiction, and perceived social acceptability. In addition, the survey also covered questions regarding characteristics of waterpipe users such as the age and place of starting the waterpipe smoking, willingness to quit it, and ownership of a waterpipe. Variable selection was guided by the health behavior models [25]. In the Health Belief Model (HBM) [25,26], perceived susceptibility, perceived severity, and perceived benefits along with mediating factors such as demographic and social variables affect an individual's decision to engage in a behavior. Thus, participant's perceived risk, worry about harm, and perceived addiction can affect whether they continue waterpipe smoking or not. According to the theory of Reasoned Action/planned behavior (TRA) [25], behavior is influenced by the intention to perform the behavior, which is affected by subjective norms and attitudes. Thus, social acceptability can affect whether a participant will continue to engage in waterpipe smoking or not.

Participants who reported having used a waterpipe to smoke tobacco previously were included in our sample for further analysis. This was based on the survey question: (have you ever tried smoking tobacco in a waterpipe (hookah, shisha, narghile, argeela), even one or two puffs?).

Two outcome variables were identified: outcome 1 was defined as: monthly waterpipe smoking or more vs. less than monthly use, outcome 2: weekly waterpipe smoking or more vs. less frequent use.

Both of these outcomes were identified by the following question: Which of the following choices best describes how often you smoke tobacco in a waterpipe? Options included: at least once a year but not monthly, at least once a month but not weekly, and at least once a week. For the first outcome, participants who answered (at least once a week) were combined with those who answered (at least monthly but not weekly).

Descriptive statistics and chi-square analyses were used to determine the frequencies and associations of participant characteristics with the two previously defined outcomes. Univariate logistic regression analyses of participant characteristics were carried out with the two outcome variables and results were presented as unadjusted odds ratios (OR) with 95% confidence intervals (CI). Two multivariate logistic regression models were carried out to determine predictors of being a persistent waterpipe user with the defined outcomes after assessing colinearity between the independent variables. Variables with probability value below 0.2 (p<0.2) in the univariate analyses were included and backward elimination was used to arrive at the final models. Gender was included in multivariate models. Results were presented as adjusted ORs with 95% CIs. All statistical analyses were carried out using SAS statistical package version 9.2.

Waterpipe smoking characteristics

A total of 2,204 students participated in the survey. Of those, 1,141 (51.77%) used a waterpipe to smoke tobacco before and were included in subsequent analyses. About one fifth of those who previously used a waterpipe (19.11%) reported smoking on a monthly basis or more frequently (model 1), and 4.57% reported smoking on a weekly basis or more frequently (model 2). The sample population characteristics are described in Table 1. Half of the sample were male (50.27%) and 45% were of white non Middle-Eastern ethnicity. Other ethnicities included Hispanic or Latino (18.24%), white Middle-Eastern descent (8.37%), and Indian Asian or Pakistani Asian (8.19%), and others (19.38%) including Black/ biracial/American Indian/Other Asian. Majority of the sample (62.01%) were 22 years old or older, and had Middle-Eastern friends (75.68%). A large portion of the sample had used cigarette (82.19%) and cigar (76.84%) before. Approximately 76% of the sample believed it was harmful to health but 57% believed it was less harmful than cigarettes. Most of the sample thought there is no or low chance to get addicted when using a waterpipe socially (67%), and the majority (83%) viewed waterpipe smoking as socially acceptable among peers. Half of the sample (49%) smoked waterpipe for less than half an hour, followed by 30-60 minutes (29%) and more than 60 minutes (22%). Most of the sample (81%) did not own a waterpipe at home.

(Table 1 Data shown as Supplementary) also summarizes the results of the chi-square analyses of various sample characteristics with the two outcome variables defined above.

Logistic regression analyses

Univariate and multivariate logistic regression results are presented in (Table 2 Data shown as Supplementary). In the multivariate models (adjusted rates), those of white Middle-Eastern descent were more likely to use a waterpipe to smoke tobacco both monthly or more frequently (model 1) (OR=3.50, 95%CI=1.63-7.49) and weekly or more frequently (model 2) (OR=9.28, 95%CI=3.03-28.41) as compared to those of white non Middle-Eastern descent, and those of Indian Asian or Pakistani Asian descent were more likely to be frequent waterpipe smokers in both of the models (OR= 2.85, 95%CI=1.33-6.13 model 1, and OR= 8.09, 95%CI=2.36-27.71 model 2). Those who smoked cigar in the past 30 days (OR= 2.24, 95%CI=1.36-3.67 model 1, and OR= 2.99, 95%CI=1.30-6.92 model 2) and had a waterpipe at home (OR= 4.16 model 1, 95%CI=2.58-6.72, and OR= 5.51, 95%CI=2.49-12.22 model 2) were more likely to use a waterpipe to smoke tobacco frequently. Those who thought government should evaluate the safety before selling the waterpipe were less likely to smoke waterpipe frequently in both of the models (OR= 0.54 model 1, 95%CI=0.33-0.88, and OR=0.29, 95%CI=0.12-0.73 model 2).

Smokers who usually smoked waterpipe for longer than 30 minute sessions were more likely to be habitual users smoking on a monthly basis or more frequently as compared to smokers who smoked less than 30 minutes (OR=2.19, 95%CI=1.26-3.79 for smoking 30-60 minutes and OR=3.25, 95%CI=1.82-5.81 for smoking more than 60 minutes for model 1). Those who felt their peers look cool or very cool while smoking waterpipe were more likely to use waterpipe frequently in the first model (OR= 2.66, 95%CI=1.56-4.54). Those who believed that waterpipe smoking was harmful to one's health were less likely to use waterpipe to smoke tobacco on a weekly basis (OR=0.30, 95% CI=0.13-0.67 model 2).

Variables that were significant in the univariate models but were no longer significant after adjusting for variables in the multivariate model include gender, age, having Middle-Eastern friends, and smoking cigarette in the past 30 days.

Among 1,141 participants who have previously tried using waterpipe to smoke tobacco, 20% became habitual users on a monthly or more basis, and 5% on a weekly or more basis. Although the population of those who tried a waterpipe was comprised of approximately 50% of each gender, habitual users comprised of a higher percentage of females. Gender, however, was not a significant predictor of habitual use in the multivariate models. This contrasts with gender differences documented in cigarette smoking where higher smoking rates were reported among males compared to females [27]. Previous studies have shown waterpipe smoking to be more appealing for females compared to cigarettes in the US [9,22] as well as in Middle Eastern countries where waterpipe smoking originated [10], as it is considered more socially acceptable for females to smoke waterpipe than cigarettes [28,31].

Our results indicate that having a Middle Eastern ethnicity or Indian or Pakistani origin is a significant predictor of habitual waterpipe smoking in both models, which might be reflective of the cultural acceptability of waterpipe smoking among these ethnic groups, as this form of smoking is also common in countries of origin of these ethnicities [10]. These results highlight the importance of developing culturally appropriate interventions within these ethnic groups to depidemiolease waterpipe smoking withinin these sub-populations which could also help prevent further dissemination of waterpipe smoking into the US society.

Past 30 days cigar use was also significant predictors of habitual waterpipe smoking in both models. Individuals who consume tobacco products may be willing to consume different forms of tobacco, and concurrent use of tobacco products could contribute to nicotine dependence, which can be maintained later by tobacco consumption by a variety of methods [22,32]. Eissenberg et al. [22] examined predictors of past 30 day waterpipe smoking among college students and found both past 30 day cigar and past 30 day cigarette use to be predictors. We did not find past cigarette use to be a predictor of habitual waterpipe smoking despite initial significance in the univariatemodel.

The majority of the student population that tried using a waterpipe in the past (75%) believed it was harmful to health, but those who thought that waterpipe smoking is harmful were 70% less likely to become habitual users in the weekly or more frequent model despite trying it in the past. In addition, participants who tried waterpipe smoking yet believed that a government agency should evaluate its safety prior to sale were also 70% less likely to become habitual users indicating a caution towards possible harms. Perceived addictiveness of waterpipe smoking, however, did not emerge as a predictor of habitual use. While Eissenberg et al. [22] reported harm perception to be a predictor of waterpipe smoking, it was not a significant predictor in the multivariate model as reported by [9]. These results indicate that perceived addictiveness might be more influential than perceived harm in the initial decision to engage in waterpipe smoking, but harm perception could be more influential in the decision to continue the habit. Consequently, educating waterpipe smokers about the expected harms could be beneficial in preventing the habit and future research should focus on designing and evaluating the possible benefits of implementing educational programs to combat the habit.

Perceived popularity and perceived social acceptability of waterpipe smoking have been reported to be strong predictors of using it in the past month [9,22]. We did not find the perceived social acceptability to be associated with continued use in this study, which may indicate that social acceptability can promote trying the waterpipe but not necessarily continued use. Nonetheless, approximately half of the population that previously tried watepipe smoking (53%) considered the habit cool or very cool, and those who believed so were more likely to continue the habit on a monthly or more basis. Future research should examine ways to modify the cool image associated with this emerging health hazard.

Students who smoked the waterpipe for more than half an hour were more likely to be habitual users in the monthly or more model. Previous reports have shown that the longer duration of the waterpipe session (>45minutes) results in a significantly greater nicotine exposure [17], which may lead to a continued use of a waterpipe to smoke tobacco. Owning a waterpipe was found to be significantly associated with habitual use in both multivariate models, as a habitual user might find it convenient to buy a waterpipe to use it at home. Maziak et al. [13] conducted a study among 268 waterpipe smokers in Syria and reported that waterpipe smoking mainly at home and alone were predictors of a higher frequency use.

Our study had some limitations. This study was based on a voluntary survey, thus, we were unable to examine nonparticipants. Our sample distribution of gender and race, however, are somewhat similar to those reported in the University of Houston Annual report for the student body [33]. Causality cannot be inferred from the study since it had across sectional design. Furthermore, generalizability of our findings may be limited to US college students and other geographic areas in the US may not be similar.

Despite these limitations, this is the first study, to our knowledge, that examines the predictors of habitual waterpipe smoking among a group of US university students who previously tried a waterpipe. Results of our study underscore the importance of developing interventions that incorporate the predictors identified. Educational programs that explain potential harms and perceived harm and modify the cool image associated with waterpipe smoking could be effective in preventing the spread of this rapidly emerging health hazard.