Objectives. We examined prevalence rates of water pipe tobacco smoking among young people as a first step in assessing the health implications of this form of tobacco use.
Methods. We examined water pipe use with data from the 2007 Florida Youth Tobacco Survey, which assessed tobacco-related beliefs, attitudes, and behaviors among the state's middle and high school students.
Results. Four percent of middle school students and 11% of high school students reported ever having used a water pipe. Adolescent boys were significantly more likely than adolescent girls to use water pipes, and African American adolescents were significantly less likely than adolescents from other racial/ethnic backgrounds to do so. Those who indicated ever having tried cigarettes and those who reported positive attitudes toward the social nature of cigarette use were more likely to have tried water pipes.
Conclusions. Water pipe use appears to be widespread among middle and high school students. Further research is needed to assess the health risks associated with water pipe tobacco smoking as well as young people's attitudes toward this form of tobacco use.
Although the results of self-report studies indicate the apparent popularity of water pipe tobacco (also known as “hookah”) smoking among adolescents and young adults,1–3 no population-based studies, to our knowledge, have estimated the prevalence of this form of tobacco use. Very little is known regarding water pipe tobacco smoking among adolescent groups. The few self-report studies that have examined use among adolescents have oversampled Arab American students, given the tradition of this form of tobacco use among individuals from the Middle East, Southeast Asia, and South Africa. For example, Weglicki et al.1 compared water pipe tobacco and cigarette smoking rates among Arab Americans (representing 45% of their sample) and non-Arab Americans. The rate of water pipe use was higher among Arab Americans than among their non–Arab American counterparts (17% vs 11%).1
In addition to assessing patterns of water pipe tobacco smoking, it is important to assess the relationships between water pipe use and use of other tobacco products. Studies have shown that adolescents with a history of water pipe tobacco smoking are 4 to 16.5 times more likely than their counterparts without such a history to experiment with cigarettes.2,4–6 Also, higher rates of water pipe use have been reported among adolescent males than among adolescent females.2,5,6
Although water pipe smoking is spreading worldwide,7 the overall prevalence of this practice in the United States remains unknown. However, estimates are available for certain groups, including college students. Smith et al.,8 in describing data collected in 2004, found that 15.3% of freshmen attending an East Coast private university had used a water pipe in the preceding 30 days. In another study, conducted in 2006 at Virginia Commonwealth University, 20.4% of students indicated water pipe use in the preceding 30 days and 48.4% reported a history of water pipe use.9
In their study of a random sample of students enrolled at a large, urban, public university in the Northeast, Primack et al.10 found that 9.5% of students had used a water pipe in the preceding 30 days, 31% had done so in the past year, and 41% had a history of water pipe use. Of those who had used a water pipe in the past year, more than one third (35.4%) had never smoked a cigarette.10 Jackson and Aveyard11 found similar results in a cross-sectional survey of college students who were customers of a water pipe cafe; most students (65%) who reported regular smoking of water pipe tobacco did not smoke cigarettes. Jackson and Aveyard also reported that rates of water pipe use increased across the college years,11 a pattern opposite to that observed with cigarette smoking.12
In addition to estimating water pipe use rates among different populations in the United States, researchers have assessed perceptions of the risks involved in this practice and the correlates of water pipe tobacco smoking. In a study of students at a private East Coast university, Smith-Simone et al.13 found that friends were the most likely source of influence on water pipe use in the following year; also, students reported that water pipe tobacco smoking was the most socially acceptable form of tobacco use among their peers and that their peers looked “cooler” when using water pipes than when smoking cigarettes or cigars.
Eissenberg et al.,9 in a cross-sectional study conducted at a large public university, found that students who were current water pipe users were more likely than students with no history of water pipe use to have smoked cigarettes, cigars, or cigarillos in the preceding 30 days; to believe that water pipe use makes their peers look cool; and to believe that water pipe use is socially acceptable among their peers. They also found, in general, that water pipe users were younger, that they were less likely to be African American, that they report lower perceived harmfulness or addictiveness of water pipe use compared with cigarette use, and that they report lower perceived social acceptability of cigarette use among peers.
In the earlier-mentioned study of Primack et al.,10 the authors reported that 33% of the college students in their sample believed that water pipe smoking is less harmful than cigarette smoking, and 52% believed that it is less addictive than cigarette smoking. In addition, 36% believed that water pipe smoking is “very socially acceptable.”
Although individuals of college age seem to be the group most vulnerable to water pipe use, high school populations are also prone to the increasing popularity of this form of tobacco use, which is primarily social in nature. Many students reach 18 years of age, the legal age to purchase and use tobacco products, while they are still high school seniors, and with the dramatic increase in water pipe cafes (between 2000 and 2004, more than 200 new hookah cafes opened for business nationwide14), these students now have establishments available to them for water pipe use.
To address gaps in the current literature on water pipe tobacco smoking among adolescents, we tested 3 hypotheses. First, older, White, male adolescents will be more likely to use water pipe tobacco than younger, non-White, female adolescents. Second, a history of having tried cigarettes will increase the likelihood of trying water pipe tobacco. Finally, adolescents who believe that cigarettes are acceptable to alleviate stress and for use in social situations will be more likely to use water pipe tobacco.
We constructed models designed to assess associations between individual-level variables and water pipe tobacco smoking via hierarchical logistic regression analyses. Demographic data were included in the first model, which tested the hypothesis that water pipe tobacco smoking would be more prevalent among older, White, adolescent boys. We then added history of cigarette use in model 2 to assess our second hypothesis. Finally, in model 3, we assessed whether attitudes toward cigarette smoking with respect to social situations and stress relief would be associated with water pipe tobacco smoking, given that its relaxing and social nature is often cited as a reason for its use.
We used data from the 2007 Florida Youth Tobacco Survey (FYTS).15 The FYTS is a statewide, anonymous, school-based survey that has been administered annually to a random sample of Florida public middle and high school students since 1998. Data are collected on students' demographic characteristics (age, gender, grade, race/ethnicity); tobacco use, tobacco sources, and quit attempts; secondhand smoke exposure; exposure to tobacco use prevention education; beliefs and attitudes about tobacco use (tendency to use in the future, health attitudes toward tobacco use, social attitudes toward tobacco use); awareness of anti-tobacco media and influence from tobacco companies; and perceptions of Florida tobacco laws.
The survey involves a 2-stage cluster probability sample design. First, a random sample of public middle schools and high schools (grades 6–12) is selected for participation in the survey. Second, within each selected school, a random sample of classrooms is selected, and students in those classrooms are invited to participate in the survey. Larger schools are sampled with greater probabilities of selection than smaller schools to ensure that every student in the state had the same probability of selection. Parental consent is mostly obtained through passive permission forms that parents must return to opt out their child from participation. In a few counties, however, active permission is required and is obtained.
In 2007, 5037 middle and 4028 high school students in 188 schools participated in the FYTS. Participation rates were calculated separately for schools and students as a ratio of number participating divided by number selected. Combined participation rates were calculated through multiplying the 2 separate school and student participation rates by each other. These combined rates were 74% for middle schools and 57% for high schools.
Students were asked 2 questions to establish race/ethnicity. The first question asked whether the student was Hispanic or Latino, and the second asked “How do you best describe yourself?” (students were instructed to select only 1 response). Response options were as follows: American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or other Pacific Islander, White, and other. Students who indicated that they were Hispanic or Latino were coded as such, regardless of their response to the second question. Students who reported that they were not Hispanic or Latino were then separated into one of 3 categories: non-Hispanic White, non-Hispanic Black, or other.
With respect to water pipe use, students were asked “Have you ever smoked tobacco out of a hookah (also called a ‘water pipe’), even one or two puffs?” If students answered yes, they were categorized as having a history of water pipe use.
As a means of establishing history of cigarette use, students were asked “Have you ever tried cigarette smoking, even one or two puffs?” Students who answered yes were categorized as having a history of cigarette use. Current cigarette use was defined with the following question: “During the past 30 days, on how many days did you smoke cigarettes?” If students answered 1 day or more, they were categorized as current cigarette users.
Two questions were used to assess young people's attitudes about why they might smoke cigarettes: “Do you think smoking cigarettes helps people feel more comfortable at parties and in other social situations?” and “Do you think smoking cigarettes helps people relieve stress?” Response options for each question were definitely yes, probably yes, probably not, and definitely not. The definitely yes and probably yes were collapsed into 1 category, and the probably not and definitely not responses were collapsed into 1 category, so there were 2 response categories for each question.
FYTS data are statistically weighted to be representative of all Florida public middle and high school students. We used SAS version 9.1 (SAS Institute Inc, Cary, NC) to establish prevalence rates of water pipe and cigarette use, analyzed separately according to gender, race/ethnicity, grade, and age. We then conducted χ2 analyses, separately for middle school and high school students, to assess the relationships between water pipe use and sociodemographic variables, other tobacco use questions, and cigarette attitude questions.
Finally, we used SAS-Callable SUDAAN version 10 (Research Triangle Institute, Research Triangle Park, NC) to conduct hierarchical logistic regression analyses assessing water pipe use. We first added the sociodemographic variables, then cigarette use responses, then attitude toward cigarette use responses to assess the influence of these predictors on water pipe use. We used the Bonferroni correction to minimize the possibility of a type I error. The overall prevalence of missing data for the 2007 FYTS was less than 5%. Cases in which data were missing for the variables of interest in this study were not included in our analyses.
Approximately 48.2% of the 8995 respondents in the 2007 FYTS were adolescent boys, 55.6% were in middle school, and 44.4% were in high school. In terms of race/ethnicity, 45.4% were non-Hispanic White, 19.0% were non-Hispanic Black, 26.3% were Hispanic, and 9.3% were from other racial/ethnic backgrounds. As mentioned, these prevalence estimates were weighted to be representative of the Florida public school system in 2007.
Overall, 4% of middle school students reported having used a water pipe to smoke tobacco, whereas 21% reported a history of cigarette use. Approximately 11% of high school students had ever tried water pipe tobacco (with more than 16% having done so before 12th grade), and nearly 41% had a history of cigarette use. Differences in water pipe use among middle school students according to sociodemographic characteristics, history of cigarette use, and attitudinal measures of cigarette use are shown in Table 1. Statistically significant differences were found for each of the associated variables such that rates of use were higher among adolescent boys, older students, those who had ever or currently smoked, and those offering positive responses to the cigarette attitudinal statements; rates were lower among non-Hispanic Black students than among students from other racial/ethnic groups.
Results of χ2 Analysis of Water Pipe Use Among Middle School Students: Florida Youth Tobacco Survey, 2007
|Students, No. (%)||χ2**|
|Boys (n = 2394)||115 (4.8)|
|Girls (n = 2488)||77 (3.1)|
|Non-Hispanic White (n = 2173)||90 (4.1)|
|Non-Hispanic Black (n = 950)||17 (1.8)|
|Hispanic (n = 1263)||58 (4.6)|
|Other (n = 488)||31 (6.4)|
|6 (n = 1743)||37 (2.1)|
|7 (n = 1535)||56 (3.6)|
|8 (n = 1619)||102 (6.3)|
|History of cigarette use||150.14|
|Yes (n = 1022)||166 (16.2)|
|No (n = 3832)||23 (0.6)|
|Current cigarette use||77.58|
|Yes (n = 285)||90 (31.6)|
|No (n = 4515)||81 (2.8)|
|Believes smoking cigarettes helps relieve stress||64.66|
|Yes (n = 2281)||145 (6.4)|
|No (n = 2483)||40 (1.6)|
|Believes smoking cigarettes makes social situations more comfortable||88.72|
|Yes (n = 1764)||144 (8.2)|
|No (n = 3026)||41 (1.4)|
**P < .05.
Differences in water pipe use among high school students are shown in Table 2. As with the middle school students, statistically significant associations were found for each of the variables such that rates of water pipe use were higher among adolescent boys, older students, those who had ever or currently smoked, and those responding positively to the cigarette attitudinal statements. Again, rates were lower among non-Hispanic Black students than among students from other racial/ethnic groups.
Results of χ2 Analysis of Water Pipe Use Among High School Students: Florida Youth Tobacco Survey, 2007
|Students, No. (%)||χ2**|
|Boys (n = 1840)||227 (12.3)|
|Girls (n = 2054)||192 (9.3)|
|Non-Hispanic White (n = 1810)||228 (12.6)|
|Non-Hispanic Black (n = 716)||22 (3.1)|
|Hispanic (n = 1043)||114 (10.9)|
|Other (n = 324)||54 (16.7)|
|9 (n = 1133)||88 (7.8)|
|10 (n = 1052)||95 (9.0)|
|11 (n = 1014)||108 (10.7)|
|12 (n = 707)||120 (17.0)|
|History of cigarette use||307.70|
|Yes (n = 1582)||355 (22.4)|
|No (n = 2280)||47 (2.1)|
|Current cigarette use||151.12|
|Yes (n = 556)||196 (35.3)|
|No (n = 3247)||181 (5.6)|
|Believes smoking cigarettes helps relieve stress||126.38|
|Yes (n = 2352)||341 (14.5)|
|No (n = 1465)||67 (4.6)|
|Believes smoking cigarettes makes social situations more comfortable||68.66|
|Yes (n = 2027)||299 (14.8)|
|No (n = 1788)||108 (6.0)|
**P < .05.
Results from the logistic regression analyses are presented in Table 3. Model 1, which included the sociodemographic variables as predictors of water pipe use, produced a good fit to the data according to the Hosmer–Lemeshow χ2 test (χ28 = 13.9; P = .09). Male students were 1.4 times more likely than were adolescent girls to use water pipes (P < .001). There were also significant differences according to race/ethnicity such that non-Hispanic Whites, Hispanics, and students from other racial/ethnic backgrounds, respectively, were 3.7, 3.4, and 5.3 times more likely to use water pipes than were non-Hispanic Blacks (P < .001). There was a trend (with 6th grade as the reference group) for increases in rates of water pipe use with each increasing grade; rates were 2.7 times higher in 8th grade than in 6th grade and 8.3 times higher in 12th grade than in 6th grade.
Results of Multiple Hierarchical Logistic Regression Analyses Predicting Water Pipe Use: Florida Youth Tobacco Survey, 2007
|Model 1,a OR (95% CI)||Model 2, OR (95% CI)||Model 3, OR (95% CI)||Model 4,a OR (95% CI)||Model 5,a OR (95% CI)|
|Boys||1.4** (1.2, 1.7)||1.5** (1.3, 1.9)||1.6** (1.3, 2.0)||1.6** (1.3, 1.9)||1.6** (1.3, 2.0)|
|Non-Hispanic Black (Ref)||1.0||1.0||1.0||1.0||1.0|
|Non-Hispanic White||3.7** (2.5, 5.4)||3.5** (2.3, 5.3)||4.1** (2.6, 6.5)||4.1** (2.6, 6.5)||4.2** (2.6, 6.7)|
|Hispanic||3.4** (2.3, 5.0)||3.4** (2.2, 5.4)||4.3** (2.7, 6.9)||4.3** (2.7, 6.9)||4.5** (2.8, 7.3)|
|Other||5.3** (3.4, 8.3)||5.1** (3.1, 8.4)||6.0** (3.5, 10.3)||5.6** (3.3, 9.6)||6.1** (3.5, 10.4)|
|7||1.4 (0.9, 2.2)||1.0 (0.6, 1.5)||0.8 (0.5, 1.4)||0.8 (0.5, 1.4)||0.8 (0.5, 1.3)|
|8||2.7** (1.8, 4.1)||1.4 (0.9, 2.1)||1.2 (0.8, 1.9)||1.2 (0.8, 1.9)||1.2 (0.7, 1.8)|
|9||3.8** (2.5, 5.8)||1.7 (1.1, 2.7)||1.5 (0.9, 2.3)||1.5 (0.9, 2.4)||1.4 (0.9, 2.2)|
|10||4.1** (2.7, 6.2)||1.6* (1.0, 2.4)||1.4 (0.9, 2.1)||1.4 (0.9, 2.2)||1.3 (0.8, 2.1)|
|11||5.4** (3.6, 8.1)||2.1** (1.3, 3.2)||1.9** (1.2, 2.9)||1.8** (1.2, 2.9)||1.8** (1.1, 2.7)|
|12||8.3** (5.5, 12.6)||3.1** (2.0, 4.9)||2.9** (1.8, 4.5)||2.9** (1.8, 4.6)||2.8** (1.8, 4.4)|
|History of cigarette use||16.3** (12.1, 22.0)||14.9** (10.9, 20.4)||15.4** (11.3, 21.0)||14.3** (10.4, 19.5)|
|Attitudes toward cigarette use|
|Helps relieve stress||2.5** (1.9, 3.3)||1.8** (1.4, 2.3)|
|Helps make social situations more comfortable||2.2** (1.7, 2.7)||2.1** (1.5, 2.7)|
Note. CI = confidence interval; OR = odds ratio. Sample size was n = 8995.
aModel significant at P < .05.
*P <.10; **P < .05.
We found partial support for our first hypothesis, with rates of water pipe tobacco use being higher among adolescent boys and older students. Although differences between racial/ethnic groups were observed, we cannot assert that White students are most likely to use water pipes given that not only these students, but also Hispanic students and students from other racial/ethnic backgrounds reported higher frequencies of use than did non-Hispanic Black students.
In model 2, we added history of cigarette use to the sociodemographic predictors of water pipe use. The results of this model showed that adolescent boys were 1.5 times more likely than were adolescent girls (P < .001) to use water pipes. After control for history of cigarette use, racial/ethnic differences continued to be similar to those observed in model 1; that is, non-Hispanic Whites, Hispanics, and students from other racial/ethnic backgrounds, respectively, were 3.5, 3.4, and 5.1 times more likely to smoke water pipes than were non-Hispanic blacks (P < .001).
The addition of history of cigarette use affected differences in water pipe use between middle and high school students. Ninth-grade students were 1.7 times more likely than were 6th-grade students to use water pipes, and 12th-grade students were 3.1 times more likely than were 6th-grade students to do so (P < .05). Students with a history of cigarette use were 16.3 times more likely than were nonsmokers to use water pipes (P < .001).
Model 2 did not produce a good fit to the data according to the Hosmer–Lemeshow χ2 test (χ28 = 27.2; P < .05). Although history of cigarette use was a predictor of water pipe tobacco use, the lack of fit for this model indicates that additional characteristics must be considered to accurately predict water pipe use among our population.
We added the attitudinal measures of cigarette use (to relieve stress and to make people comfortable in social situations) in models 3 and 4. Similar to the previous models, gender and race/ethnicity remained significant predictors of water pipe tobacco smoking. Grade-level differences did not appear until grades 11 and 12. A history of cigarette use was a strong predictor of water pipe use (those with a history of cigarette use were approximately 15 times more likely to use water pipes than were those without such a history).
The results for model 3, which included the stress attitudinal variable, indicated that students who reported positive (versus negative) feelings regarding use of cigarettes to relieve stress were 2.6 times more likely to have used a water pipe (P < .001). However, this model was not a good fit to the data (χ28 = 19.0; P < .05). Students who reported positive (versus negative) feelings regarding the use of cigarettes in social situations (model 4) were 2.2 times more likely to have used a water pipe (P < .001). The Hosmer–Lemeshow χ2 test for model fit indicated that model 4 produced a good fit to the data (χ28 = 9.3; P = .3).
Model 5, the full model, included all of the variables from the bivariate analyses. Patterns similar to those observed in the previous models remained. Adolescent boys were 1.6 times more likely than were adolescent girls to use water pipes. In comparison with non-Hispanic Blacks, non-Hispanic Whites and Hispanics were more than 4 times as likely to use water pipes, whereas students from other racial/ethnic backgrounds were more than 6 times as likely to do so.
Grade-specific differences did not appear until the upper grades, with 11th and 12th graders being approximately 2 times more likely than 6th graders to use water pipes. Those reporting that they had tried cigarettes were 14.3 times more likely to use water pipes than were those who had not tried cigarettes. Students with positive (versus negative) attitudes toward smoking cigarettes to relieve stress were 1.8 times more likely to use water pipes, and students with positive (versus negative) attitudes toward smoking cigarettes in social situations were 2.1 times more likely to do so.
The Hosmer–Lemeshow χ2 test for model fit indicated that model 5 produced a good fit to the data (χ28 = 10.2; P = .2). Therefore, we found support for our third hypothesis, according to which positive attitudes toward cigarette smoking to relieve stress and to make people comfortable in social situations would predict water pipe use.
Our results suggest that water pipe tobacco smoking is prevalent among high school students. The overall rate among high students taking part in the 2007 FYTS was 11% and the rate among 12th-grade students was 16%, comparable to the prevalence found among 9th- through 12th-grade students in a midwestern state (6.0%–17.0%).1 Water pipe use was low among 6th- through 8th-grade students, at 4% overall.
In general, rates of water pipe use were higher among adolescent boys than among adolescent girls, higher among White and Hispanic students and students from other racial/ethnic backgrounds than among Black students, and higher among 11th- and 12th-grade students than among students in lower grades. In each of our models (with non-Hispanic Blacks as the reference group), rates were highest among students in the “other” racial/ethnic category. Given that Florida has the fourth-highest population of Arab Americans in the United States16 and that the FYTS does not report specific data for Arab American students, we can speculate that the significant number of such students is 1 reason for this result. Water pipes have a long history of use among those of Arabic descent.
In one of our models (model 2), we explored the relationship between history of water pipe use and history of cigarette use and found that cigarette smoking was associated with water pipe tobacco smoking. Unfortunately, we cannot determine the direction of this relationship from our cross-sectional analysis. Also, the model did not show a good fit, indicating that although there is a correlation between trying cigarettes and trying water pipes, variables in addition to demographic characteristics and tobacco experimentation need to be included in models attempting to explain this relationship.
We also assessed the relationship between water pipe tobacco smoking and attitudes toward cigarette smoking. Although the questions addressing attitudes were specific to cigarette use as opposed to water pipe use, they are still helpful in assessing why an adolescent might be interested in using any tobacco product. We focused on 2 attitudinal measures that shared a direct relationship with reasons cited for using water pipes in other studies9,10: to feel comfortable in social situations and to reduce stress. Positive responses to these attitudinal statements were associated with a history of water pipe use.
Given that the model including the social situation measure produced a good fit to the data, whereas the model including the stress measure did not, we would revise our hypothesis to include only the acceptability of smoking in social situations as a predictor of water pipe use. This notion fits with research indicating that water pipe tobacco smoking is a social activity.9,10 Smith-Simone et al.3 found that water pipe use is more common on weekends, another indication of the social nature of this particular form of tobacco use. The rising numbers of hookah (water pipe) cafes in recent years present young people with increasing opportunities to engage in a behavior that seemingly is socially sanctioned.
We postulate that water pipe use is on the rise among tobacco experimenters who are seeking a social outlet. The high prevalence rates among adolescents, high school students, and college students perhaps indicate that water pipe users do not self-identify as smokers or that they do not recognize water pipe use as detrimental to their health.3,8
Some limitations of our study should be noted. For example, we assessed only whether respondents had ever used a water pipe to smoke tobacco; we did not determine whether they were currently using water pipes on a regular or semiregular basis. Future studies should assess frequency of use as well as history of use.17 Also, because the question focusing on water pipe use was introduced in the FYTS for the first time in 2007, we cannot measure the direction of the relationship between ever having tried cigarettes and ever having tried a water pipe. Future research should focus on the direction of this relationship as well as on descriptive profiles of individuals who might choose to smoke only water pipes or to smoke only cigarettes.
In addition, although the FYTS is supported by the Florida Department of Health, it is limited to public schools. Therefore, generalizations can be made to students in Florida public schools but not to students in private schools or those who drop out of school. Also, the participation rates (74% for middle schools and 57% for high schools) were low for school-based surveys, which could have biased our results. However, given the likelihood that refusal rates are higher among tobacco users, our results may underestimate water pipe use among adolescents in Florida. Because of the somewhat high levels of use reported among Florida adolescents, there is a need for education and intervention programs as well as further investigation of prevalence rates in the state.
As mentioned, to our knowledge no other studies of the prevalence of water pipe use among middle and high school students have involved population-based surveys. Future research is needed to determine prevalence rates among not only middle and high school students but also college students and young professionals. In addition to prevalence rates, patterns and health risks of water pipe use require further investigation. Both the social nature of this form of tobacco use and false beliefs about the safety of water pipe tobacco warrant examination.
We propose that water pipe use is widespread and is not a fad. Water pipe users are becoming younger, and water pipe cafes are opening at alarming rates. In addition to the dangers of the heated tobacco and accompanying toxins, the sharing of water pipes may lead to an increase in orally transmitted infectious diseases. There is a need for further research and education surrounding water pipe use, and public health personnel can play an important role in enhancing knowledge in this area.
Human Participant Protection
This study was approved by the University of Florida institutional review board.