Prevalence, Predictors, and Characteristics of Waterpipe Smoking Among Jazan University Students in Saudi Arabia: A Cross-Sectional Study

Background: Waterpipe smoking (WPS), also called shisha, has considerable short and long-term effects on human health. WPS has become increasingly popular among the youth in Jazan society. Hence, this study is aimed to assess the prevalence, predictors, and characteristics (knowledge and attitude) of WPS among male and female students at Jazan University. Methods and materials: A descriptive, cross-sectional study was conducted among 405 students. Data were collected using a self-administered validated questionnaire. The prevalence and predictors of WPS and the association between important socio-demographic factors (sex, type of college, parents smoking, home mate smoking and close friend smoking) were studied. Results: The prevalence of WPS among students was high, approximately 34.0%. The prevalence rate was significantly higher in males (42.5%, n = 74) than in females (27.0%, n = 57) (p = 0.001). The main predictors of WPS were: being male (OR = 1.99, 95% CI [1.30, 3.06], p = 0.001), believing that shisha smoking is less harmful & addictive than cigarettes (OR = 3.84, 95% CI [1.88, 7.83], p < 0.001 and 3.80, 95% CI [2.0, 7.11], p < 0.001, respectively), and having a close friend who smokes (OR = 6.85, 95% CI [3.84, 12.22], p < 0.001). Conclusions: WPS prevalence among Jazan University students was high, and the most influential factors were being male, having smoker housemates and friends, and having incorrect thoughts and beliefs.

respectively [15]. In Karachi, Pakistan (2009), the prevalence among 422 college students was 45.2% [1]. In the United Kingdom a 2015 study was done to measure the prevalence and patterns of water-pipe tobacco use among 2217 student from six universities, it found a total of 66.0% (n = 1409) reported having ever tried water-pipe tobacco, and 14.4% (n = 300) reported past-30 day waterpipe tobacco use [16].
Numerous people wrongly believe that WPS is less harmful than cigarette smoking [2,3,9]. Also, the research team believes that WPS is gaining popularity among younger generations, among females and in social gatherings. Waterpipe smoking is being offered in many restaurants and coffee shops in the Jazan region. Therefore, this study focuses on the prevalence of WPS and its predictors among Jazan University students which were not assessed by previous studies.

Study design, settings, and population
This study is an observational, descriptive, cross-sectional, and web-based survey to assess the prevalence of WPS and its determinants among male and female students in Jazan University. The inclusion criteria were students at Jazan University enrolled in the academic year 2018. Excluded from the study were visiting students from outside Jazan University.

Sample size and sampling procedures
Using a multistage stratified simple random sampling, a study sample from Jazan University students was selected. First, the colleges were divided into two strata (medical and nonmedical) and then the colleges were randomly selected by lottery method from each stratum according to their number. Three colleges were selected from the medical stratum (Medicine, Applied Medical Science and Pharmacy) and another three colleges were selected from the nonmedical stratum (Science, Computer Sciences and Information Systems and Education).
The calculated sample size was divided between these six colleges proportionate to the size of students in these programs. Participating students within these colleges were selected by simple random technique, and they received a web-based survey via social networks. They used their own internet to fill out the questionnaire and at a time and place of their convenience.
The calculated sample size for the study was 405 participants, and it was calculated by the following equation for a cross-sectional study: n = z 2 × (p) × (1p − p)/d 2 , where the 95% CI, the anticipated population proportion (p) of the sample, was estimated to be 50%. This team felt 50% was the safest choice for p because the required sample size is largest when p = 50% and when the required absolute precision on either side of the anticipated population proportion (in percentage points) is d = 5% (0.05). An additional 5% was allotted for the anticipated non-response rate.

Data collection and quality control
Data were collected using a 45-item self-administered questionnaire, which had been used in previous research on the same population [17], and we obtained the necessary permission to use it. The questionnaire was divided into three parts. The first part contains five-item demographic and background data. The second part contains 11 items on respondents' knowledge and attitude regarding WPS, its effect on health, and knowledge about tobacco and nicotine (which are present in WPS). The third part, which includes 29 items, pertains to the practice of frequent WPS and the reasons that lead the participant to use it. The questionnaire was revised by some experts in the epidemiology and behavioral and social science departments to evaluate the validity of its content according to a previous study by Bahri et al. [17] In addition, face validity was assessed by piloting the questionnaire on a sample of 30 students to ensure clarity and to guarantee that all participants would be able to answer the questionnaire completely.

Data management and analysis
The collected data were verified manually, encoded to a personal computer, and then analysed using the Statistical Package for the Social Sciences version 22. Descriptive statistics were calculated for study variables -e.g., frequency and percentage for qualitative variables, and mean and standard deviation for quantitative variables. Appropriate significance tests (such as chi-square) were applied. Binary logistic regression analysis was performed to detect the predictors of WPS; p < 0.05 was considered to be statistically significant.

Ethical consideration
Ethical approval was obtained from the Institutional Review Board of Jazan University, Kingdom of Saudi Arabia. Data privacy and confidentiality were maintained throughout the project. Furthermore, informed consent was a prerequisite to complete the online survey.

Socio-demographic data of included participants
A total of 385 students responded to the questionnaire (response rate 95%). The mean participant age was 22.03 years ( Table 1). Most participants (54.8%) were female, with males constituting only 45.2%. The majority of students (94, 24.4%) were from the Faculty of Applied Medical Sciences followed by the Faculty of Education (87, 22.6%) and the Faculty of Medicine (74, 19.2%). Regarding residence, more students (50.9%) were from rural areas than from urban areas (49.1%). Moreover, 74 (19.2%) students had one or both of their parents who smoked, 137 (35.5%) students had a housemate who smoked and 245 (63.6%) students had a close friend who smoked.

Knowledge and attitude towards WPS
Regarding attitude, 6.2% of participants smoke shisha 1-3 times weekly, and 2.3% smoke 4-6 times weekly. The most preferred place for smoking is with gatherings of friends (13.8%). Only 5.7% think that smoking shisha makes them look good. About 24% of participants smoke for experience and curiosity, 11.7% because of friends, and 9.4% due to boredom and emptiness. Five percent of the participants smoke cigarettes as well as smoking shisha. 7% of respondents chew qat while smoking shisha ( Table 3).
The majority of participants know that WPS causes damage to health (94.8%). About 70% of the participants know that shisha contains nicotine as a cigarette does. Only 58.7% know that shisha contains tobacco as a cigarette does. Approximately 85% of the participants thought that sharing shisha between people could convey diseases. However, only 40% thought that smoking shisha would be less addictive than cigarettes smoking. Half the participants thought that smoking shisha in different flavors, like apple, makes it less harmful than flavorless ( Table 4).

Predictors of WPS
Binary logistic regression analysis demonstrated that males were more likely to smoke shisha than females (Odd Ratio, OR = 1.99, 95% CI [1.30, 3.06], p = 0.001). Interestingly, students who thought that shisha was less harmful and less addictive than cigarettes were more likely to smoke shisha ( 4.05], p < 0.001, respectively). Therefore, smokers were 22 times more likely to accept their friends' invitations to smoke than those who had never smoked shisha (OR = 22.12, 95% CI [12.12, 40.37], p < 0.001) ( Table 4).

Discussion
This study aimed to assess the prevalence and predictors of WPS among male and female students at Jazan University. Our findings demonstrated a high prevalence of WPS among university students (34%). The main predictors of WPS were: Being male, the belief that shisha smoking is less harmful & addictive than cigarettes, the belief that the presence of the filter mechanism can protect them, and having a roommate or close friend who smokes. These findings highlight the importance of awareness campaigns to correct misconceptions. In addition, these findings clearly present the significant impact of close friends and roommates on behavior. In the case of cigarette smoking, the influence of the parents in paramount. With WPS, on the other hand, the influence of close friends and roommates is the main factor after the desire and curiosity to try shisha. This may be because shisha is often used in gatherings of friends where it is exchanged between them.
In term of knowledge and attitude, our findings showed that the knowledge of the harm created by WPS was higher in females compared to males (p = 0.032), while    males were more likely to be aware with the contents of the shisha (p = 0.003). Females preferred smoking at home and closed places more than males (p = 0.001). This can be explained by the local oriental culture, which prevents girls from smoking in public places. For males, the main causes that made them smoke shisha for the first time were friends and family, while in females, it was the experience and curiosity. Moreover, males were more likely to quit smoking than females (p = 0.020).
Our findings were supported with the previous findings of other studies conducted in Riyadh, Kingdom of Saudi Arabia in 2011 [2]. In one of these previous studies, the prevalence of WPS was high among male and female high school students, which showed that 414 (33%) of the 1272 participants had tried WPS. The current study shows that the result of prevalence was higher than that of the study conducted by Mahfouz et al. in Jazan University (2014), which showed that 14.6% of both male and female students used WPS [12].
In terms of the predictors of WPS, Bashirian et al. demonstrated that the belief in an individual's ability, support from friends, and the benefits of reducing waterpipe smoking are the most important factors in reducing WPS among students [18]. This supports our observation that friends have a significant impact in whether one will start or quit smoking. Another study by Ziaei and his colleagues reported a significant association between WPS usage and the initial offer of WPS by a close friend (OR = 3.31; 95% CI [2.17, 5.04]) [19]. Moreover, they found that females were associated with lower risk of being WP smokers (OR = 0.45; 95% CI [0.30, 0.70]). These findings were in agreement with ours. Saravanan et al., showed that social interaction is one of the major independent predictors of shisha smoking in comparison to parental smoking behavior. They explained that, due to the religious and legitimate restriction of other forms of substance abuse, students tended towards shisha smoking when they were involved in social interactions in cafeterias and other places [20].
Regarding the effect of family and friends behaviors on encouraging students to practise WPS, 46.9% of the students who had a close smoker friend were frequent WPS users, but 53.1% were non-smokers (p = 0.000). In addition, 37.8% of students who have smoking parents were also smokers; nevertheless, 62.2% of them were nonsmokers (p = 0.441). A large number of WPS students have family members at home who smoke, apart from their parents, with a significant association (66,48.2%, P = 0.000), in agreement with the 2014 study of Mahfouz et al. in Jazan [12]. This finding was also consistent with that of Hassan et al. in Riyadh, Saudi Arabia, in 2014, with the participation of Al-Ghad International College students [13]. This previous study found that 18 (48.6%, P = 0.003) of the smoking students stated that most of their friends smoked, 27 (38.0%, P = 0.000) stated that some of their friends smoked, and 16 (44.4%, P=0.7) stated that their fathers smoked. This study also agrees with the results of Al Moamary et al. conducted in Riyadh, Saudi Arabia, in 2011 [2]. They found that the prevalence of WPS among students' friends was 284 (23.15%, P < 0.001), with a strong relationship existing between WPS students and family members at home, other than their parents, who smoked 132 (10.86%, P < 0.0001) and who encouraged others to use WPS; these findings are in agreement with the current study.
Regarding the place of residence, the prevalence of WPS among students who lived in rural areas (72, 36.7%) was higher than that in urban areas (59, 31.2%), with no significant association (P = 0.253). The previous study conducted in Aleppo, Syria, in 2004 found the same frequency of WPS students in both rural and urban areas (43, 16.0% urban; 43, 13.5% rural) [15].
Moreover, the current study shows that nonmedical students smoked (71, 37%) in greater numbers than medical students who smoked (60, 31.1%), with no significant association (p = 0.223). These frequencies agree with those of the study conducted in Aleppo, Syria, in 2004, in which WPS prevalence was higher among students in science-related areas (50, 17.7%) and arts, law and humanities (24, 13.0%) than among those in medical areas (12,10.2%) [15]. These findings also agree with the study of Mahfouz et al. in Jazan in 2014 [12].

Limitations
This study has limitations. First, it acquired a sample from university students rather than the whole population of Jazan City. Second, it used a web-based questionnaire which might not be accessible to all students. Thus, students who use the Internet might have more information about the use and harmful effects of WPS than those who do not. Additionally, the study results cannot be generalised to all residents of the Jazan region.

Conclusion and recommendations
In conclusion, the prevalence of WPS among college students at Jazan University was high. The influence of siblings and close friends was clearly observed and should be considered in future awareness campaigns. The research team strongly recommends further investigation into this problem, considering that WPS substantially affects one's health. In addition, they recommend further analysis of the mechanism of WPS to one's health and the clinical diagnosis and management related to health affected by this lifestyle practice.