Evidence of emerging hookah use among university students: a cross-sectional comparison between hookah and cigarette use

Sample

We intercepted potential student participants at three locations on the UF campus. We selected locations that reached a diverse group of students in terms of major/affiliated college, sex, race, and level in school (freshman to graduate student). Three locations were used approximately 10 different days for 4–6 hours at a time. Each location’s days and times were different week to week. At the conclusion of each day of data collection, the sample was analyzed to ensure appropriate representation of the target population, with a plan to oversample if it appeared that groups were under-represented in the sample.

To the best of their ability, the research assistants approached every 10^th person that passed near the area. Access to each location was such that large groups of people would typically not pass by at the same time. For instance, at one location there is a sidewalk that will comfortably accommodate no more than two or three people side by side. We chose this sampling method as a way to force the research assistants not to select respondents based on any visual characteristics, such as their demeanor. Once they reached 75–100 completed surveys, the team moved to another location. The final sample included 1,203 participants.

Procedure

At each location, we set up five stations with laptop computers. Each computer had the survey instrument loaded into the computer-assisted personal interview (CAPI) program. Research assistants approached prospective respondents and offered them a $5 gift card for their participation. Students who agreed were given a card with a unique identifier; they then went to a station with computers, read and signed a consent form, typed in their identifier code, and handed the card to the research assistant overseeing the stations. Participants were given the gift card after completing the survey. Using this approach, multiple surveys were conducted simultaneously. The study protocol and survey instrument were approved by the UF IRB. Data were collected during the months of November and December, 2009.

Survey instrument and measures

We developed a computer-based survey instrument to collect information on study participants’ tobacco-related behaviors and demographic characteristics. A thorough literature review of past studies informed the questions included and the instrument was pilot-tested on a group of students prior to administration for the study.

Participants answered demographic questions regarding their sex, race/ethnicity, and age; academic characteristics such as current level in school (undergraduate/graduate), affiliated college (a drop-down menu of all colleges in the university), and full-time or part-time enrollment status. If individuals were not current UF students, they were not included in the study. Although ethnicity and race were asked separately on the survey instrument, when matching back to university registrar data for weighting to ensure representativeness we found that those items were not reported separately in the database. Therefore, for weighting purposes we first ensured that the proportion of Hispanic and non-Hispanic respondents matched the university registrar and then we categorized by students’ race if they were not of Hispanic ethnicity. Furthermore, those identifying as American Indian and Multiracial were combined into the Other race group for statistical testing.

Participants were asked to identify if they had ever used cigarettes or hookah/waterpipe in their lifetime. Participants who reported ever use were then asked separately whether they had used the product in the past 12 months (year) or past 30 days. After identifying which tobacco products participants had used, participants were also asked which product they tried first and the age of first use for each product. Respondents who reported hookah use were also asked to describe the social setting for use, whether they smoked with others or alone, and where they typically smoked hookah.

Data analysis

To minimize sampling bias, we computed a sampling weight for each student based on ethnicity and race, sex, level of education, and college affiliation. The sampling weight was computed based on the population distribution from the data obtained from the university registrar office for Fall, 2009. The sampling weight for each participant was calculated by dividing the race-ethnicity/sex/academic proportion for that individual in the target population by the corresponding proportion in the sample. Although race and Hispanic ethnicity were measured separately, they were combined for analytic purposes.

Chi-squared tests were used to test the associations between categorical demographic variables and both cigarette and hookah use. Two independent-sample t-tests were used to test the difference in mean age between cigarette and hookah users and non-users. Using logistic regression, odds ratios for hookah use by demographics and cigarette use were calculated. The models developed predicted hookah use at all time frames (ever, past year, or current) first with just demographic variables included (Model 1), and then with the same time frame for reported cigarette use (Model 2). These tests were calculated using SAS statistical software v9.2 (Cary, NC).

Prevalence

There was no age difference for ever use, but past year users of hookah were younger (M = 21.0) than those who had not used in the past year (M = 22.5) (t(1193) = −4.52, p < .001). This relationship was also true for current hookah use: current hookah users were slightly younger (M = 22.0) than non-current hookah users (M = 22.2; t(1191)= −3.35, p < .001). Ever users of cigarettes were slightly older than never users (22.63 vs. 21. 71; t(1202) = 2.17, p < .05) but no age differences were found for reported past year or current use.

Age at first use

Among those who reported ever using cigarettes or hookah, the average age of first tobacco use was 16.4 years (SD = 0.16). Cigarettes were more likely to be the first tobacco product tried (58.0%) than was hookah (18.9%). Respondents who reported cigarettes as their first form of tobacco use were more likely to be under the age of 18 at the time of first use (67.9% vs. 43.6%; χ ² (1) = 13.26, p<.001). Those who reported trying cigarettes first were more likely to be females (65.3%) than males (51.5%; χ ² (1) = 6.06, p < .05). In contrast, those who tried hookah first were more likely to be over 18 at first use (31.2% vs. 10.7%; χ ² (1) = 21.86, p<.0001). There were no sex differences for respondents who reported trying hookah as their first tobacco product.

Social setting

Among hookah smokers, the majority (73%) reported only smoking with friends/others and an additional 22% reported usually smoking with friends/others. Only 4% reported usually smoking alone while no one reported only smoking when alone. With regard to location, nearly all (90.2%) reported smoking in a restaurant/café/bar; 85.4% reported also smoking in a friend’s home/apartment; 51.4% reported smoking in their own home/apartment or dormitory; and 7.3% reported smoking in their parents’ home/apartment.

Hookah use predictors

Regarding dual use, for current cigarette users 38.2% were also current hookah users and slightly fewer current hookah users (35.2%) also reported current cigarette use. Among current hookah users more than one-quarter (28.8%) reported never having smoked a cigarette.

Prevalence

Other studies [3, 4, 7, 8] have reported similar rates of hookah use, but hookah smoking lagged behind cigarette smoking in those studies. The overall prevalence for ever use (46.4%), past year (28.4%), and current use (9.9%) are similar to those reported for a University sample in 2008 by Primack et al. [3], who reported ever use at 40.5%, past year use of 30.6% and current use of 9.5%. Grekin and Ayna [2], however, found much lower rates in a university sample with only 15% ever use and 12.4% past year use (this study did not report current use). Eissenberg et al. [4] reported higher rates (ever 48.4%, past year 43.4%, and current 20%) in their college study of Introduction to Psychology students. However, the cigarette rates from the same sample [4] were significantly higher than their hookah rates and they were higher than the national averages for cigarette use with ever cigarette use at 73%, past year use at 57.7% and current cigarette use at 41.5%. Finally, in a sample of eight universities, Sutfin et al. [7] also reported similar ever use prevalence of 40.3%, but a much higher current use of 17.4%. All of these studies indicate popularity of hookah, while the current study shows it may even be more popular than cigarettes.

Though prior studies [1–5, 8] found higher rates of hookah use among males, in the present study females and males reported similar rates of hookah use. This finding held even when covariates were added into a logistic regression model to predict hookah use, controlling for age and race differences. Similar trends were found for cigarette use. Again, prior research indicated higher prevalence of cigarette use for males [8, 11] across all levels of use, while this study found no sex differences in cigarette smoking rates for ever or past year use, although males surpassed females for current smoking. This is a potential trend to monitor, or there may be something unique with this sample given the high rates of hookah use for females as well.

The racial differences in both cigarette and hookah use in our sample also appeared different from past studies. We found that all levels of hookah and cigarette use (ever, past year, and current use) were most prevalent among Hispanics, followed by whites, Asians, then other. Past research on both adolescent [14] and college samples [15] indicated the highest tobacco prevalence among non-Hispanic whites, followed by Hispanics and lastly African Americans. Similar to the sex differences found in the current study, the higher hookah and cigarette use prevalence among Hispanic participants could be a reflection of this university sample which draws heavily upon students from the Miami area. Although many studies are only able to statistically assess white vs. other race (combining minority respondents), hookah use is typically reported more often for whites [4, 5, 8, 16]. Sutfin et al. [7] found no difference between white and Hispanic hookah prevalence in their sample. These inconsistent trends need to be further assessed in future studies, with enough sampling of minorities to provide detailed information.

Age

Hookah use was more likely to be reported by younger students. This pattern is also seen in research by Primack et al. [3]. In the present study, current and past year users were found to be younger compared to those who reported not using in the past month or year. The fact that individuals only need to be 18 years of age to enter the relatively large number of hookah venues available in this university town may be influential in this regard.

Age at first use findings indicated differences between cigarette and hookah users as well. Participants who reported trying cigarettes as their first tobacco product were more likely to be under age 18 at first use, while those who reported trying hookah first were more likely to be over age 18 at first use. While tobacco can be purchased and used at age 18 legally, this pattern is of concern as traditionally it was thought that not trying tobacco products by the age of 18 reduced the odds that a young adult would ever become a tobacco user [17]. The pattern of hookah use and its acceptability among these novice users may change the rates of multiple tobacco products used later into young adulthood. Cigarette users were also highly predictive of hookah use, indicating that many young adults already using cigarettes are now using at least two forms of tobacco.

Social setting

Hookah smoking is a social activity for many adolescent and young adult users. This fact held true in the current study in which the overwhelming majority reported using hookah while with others, which is supportive of previous research regarding the situation in which respondents use hookah [18]. Also supportive of past research [16, 18], participants from the current sample indicated cafes/restaurants/bars were the most common place to use hookah, followed by at friends’ houses, further supporting the social nature of this form of tobacco use.

Predictors of hookah use

There is concern that concurrent tobacco product use might potentially multiply tobacco’s deleterious effects. In this study, 35% of current hookah users also were current cigarette smokers as compared with the 55% found in other recent research [7]. Sterling and Mermelstein [1] reported that adolescent current cigarette users also report a high prevalence of hookah use, and trying cigarettes is associated with using a variety of other tobacco products. In this study, cigarette use was highly predictive of hookah used in the same time frame, even after controlling for demographic characteristics. Surveillance systems need to continue to monitor all tobacco products, including adding hookah and other alternative forms to studies on adolescent cigarette use, to understand the role certain tobacco products play in influencing multiple forms of tobacco use.

While the relationship between cigarette and hookah use is of concern, it is also important to note that of the hookah users, 65% reported not being current cigarette users. Over one-fourth (28.8%) of current hookah users reported never having tried a cigarette. This is particularly interesting as it indicates that a substantial number of the participants may have been nicotine naïve had it not been for hookah. Multiple interventions are needed to target dual users, but interventions also need to target the potential myths adopted by hookah users regarding its relative safety, even if they understand the risks in cigarette smoking [5].

Acceptance and availability

Sutfin et al. [7] demonstrated that the number of hookah venues in a geographical area was related to hookah smoking prevalence, which potentially holds true for this study as well. At the time of data collection, the community surrounding the University of Florida had seven establishments that offered hookah smoking as part business. While there is no comparison community included in this study, college towns appear to be a major target market areas for hookah venues. For example, Primack et al. [19] reported four hookah venues opening within five miles of Carnegie Mellon University and the University of Pittsburgh. Primack et al. [20] also mapped the concentration of 144 U.S. hookah establishments that advertised on the internet. They found that while those establishments were located in all regions of the country, the heaviest concentrations per population were in 11 states, including Florida. Smith et al. [16] reported that adolescents first learned about hookah from friends (50.3%) but also from seeing a hookah lounge (20.9%). The availability and acceptance of this social arena to share tobacco smoking is one possible explanation for the dramatic increase in prevalence observed among adolescents and young adults.

Limitations

The intercept sampling method remains a form of convenience sampling; however, every 10^th student was approached in the effort to reduce systematic bias. Additionally, the data were then weighted to the registrar’s list of currently enrolled students in the same academic semester to further create a more representative sample of the general student population. The prevalence results presented here are also only representative of one university in the southeast region of the United States. In general, the prevalence and patterns of tobacco use vary by locale. Similarly, states differ on policy restrictions for tobacco use in work places, public spaces, restaurants, and bars. However, the present information is helpful for understanding the use of traditional and emerging tobacco products at a major U.S. university.