Introduction
Smoking tobacco poses a serious threat to global health. According to the World Health Organization, about 8 million people worldwide die each year from smoking. In 2012, 3.1 billion people over the age of 15 in the world had smoked tobacco, and if the prevalence of tobacco smoking remains stable, this figure will almost double by 2025 [1]. Over the past three decades, the rate of tobacco smoking has decreased significantly in industrialized countries due to increasing awareness of its side effects and effective control policies, while there has been an increase in its prevalence in developing countries [2]. More than 80% of smokers live in low- and middle-income countries. The tobacco industry in these countries mainly targets young people and women. Although in many countries, women do not traditionally smoke, the target of current tobacco ads is mainly women. From 2005 to 2025, the number of women smokers is expected to double, especially in low- and middle-income countries. In 2018, 3.5% of women and 24.5% of men aged >15 years smoked tobacco In Iran [1]. 
Tobacco smoking has various types in Iran, including cigarette smoking and hookah smoking. Adolescents are more likely to smoke and are less aware of the consequences of their behavior, and less likely to quit smoking [3]. The results of a study on high school students in Tabriz City, Iran, showed that 5.9% of students smoke cigarettes, and 5% hookah regularly [4]. In another study on medical students in southern Iran, most students reported their first experience of smoking tobacco at the age of 15-20 years, and the majority had hookah smoking with their friends [5]. The main reasons for smoking hookah and cigarettes were entertainment and fun. The results of a study on students of Tehran University of Medical Sciences, Tehran City, Iran, showed that 29% smoked hookah. Also, the most important factors in the tendency to smoke hookah were filling leisure time, being affordable, reducing anxiety, relieving fatigue, being accessible, and having hookah user friends [6].
In identifying the predictors of smoking behavior, factors such as attitudes, subjective norms, intentions, and perceived behavioral control can be influential [7]. To investigate these factors, health education and behavior change models are very helpful [8, 9]. These factors are constructs of the theory of planned behavior (TPB), which seems to be a suitable model for a better understanding the predictors of smoking behavior [10]. This theory is one of the theories of behavior change. It states that intention is the main determinant of human behavior. It is a mental and social situation and creates a link between the individual and her/his action, which is under the influence of three independent constructs of attitude, subjective norms, and perceived behavioral control. 
Attitude is a person’s positive or negative evaluation of behavior. Subjective norms suggest that perceived social pressures may or may not cause a person to engage in certain behaviors. Finally, perceived behavioral control is the perceived ease/difficulty of successfully performing a particular behavior and is thought to, directly and indirectly, affect the behavior. Individuals evaluate a behavior positively and intend to do it when they believe that important, influential people in their life approve the behavior and that the behavior is under their control [9]. 
Regarding that the theories of behavior change are variables based on the cultural and social context of society, we found that few studies have been conducted to identify the factors affecting smoking behavior in adolescent girls living in Guilan Province of Iran. Thus, we aimed to determine the predictors of smoking based on the TPB in female high school students in Rasht City, Iran.

Materials and Methods
This cross-sectional study was performed on female high school students aged 14-18 years in Rasht, Iran. Based on the study by Najafi et al. [11], who reported the prevalence of smoking in students as 18%, considering a precision of 5% and a confidence level of 95%, the sample size was calculated 340. The research samples were selected with a multi-stage sampling technique. In the first stage, from each of the two education districts of Rasht city, 5 clusters were randomly selected. Then, proportional to the population of the two districts, 53% of the samples were selected from one district and 47% from the other one. In each of the schools, 20% of the sample size related to each district was selected by systematic sampling method using the list of students in the school management office. The inclusion criteria were being a female high school student in Rasht and having a willingness to participate in the study. Sampling was done from October to December 2016. By referring to the selected schools and coordinating with the school principal, one classroom was selected in each grade. From each classroom, 12 students were selected randomly. If a selected student wanted to leave the study, she would be replaced with another eligible one in the classroom.
The data collection tools were a demographic form and the TPB questionnaire for tobacco use developed by Karimy et al. [12] in Persian. In this questionnaire, 14 items are related to the construct of attitude toward the behavior scored on a 5-point Likert scale (from 1=strongly agree to 5=strongly disagree) except for items 7, 8, and 12, which have reverse scoring. Also, 7 items are related to the construct of perceived behavioral control scored on a 5-point Likert scale (from 1=extremely likely to 5=extremely unlikely) except items 5, 6, and 7, which have reverse scoring. Another 5 items are related to the construct of subjective norm scored on a 5-point Likert scale (from 1=strongly agree to 5=strongly disagree) with a total score ranging from 5 to 25. Finally, 7 items are related to intention to perform the behavior scored on a 4-point Likert-type scale (from 1=I definitely will to 4=I definitely will not) with a total score ranging 7-28.
Each student should have completed the questionnaire individually in the absence of school staff. The collected data were analyzed in SPSS v. 21. Descriptive statistics (Mean±SD, percentage) were used for describing data. The normality of data distribution was assessed with the Kolmogorov-Smirnov test. The Mann-Whitney U test was used to compare the construct scores of the TPB questionnaire for smoking. Logistic regression analysis was used to determine the matched odds ratio at a 95% confidence interval for each variable in predicting smoking behavior, and the Spearman correlation test was used to determine the correlation between the TPB constructs. The significance level was set at 0.05.

Results
About 12% of the students answered yes to the question of “Have you ever tried smoking cigarettes?”. However, 1.5% answered yes to the question of “Are you currently smoking cigarettes?”. About 41% answered yes to the question of “Have you ever tried smoking hookah?”. But 5% answered yes to the question “Have you smoked cigarettes in the past three months?”. The Mean±SD score of attitude construct was 57.2±8.1 out of 70. The Mean±SD score of the perceived behavioral control construct was 29.3±3.4 out of 35. The Mean±SD score of subjective norms was 23±2.3 out of 25, and the Mean±SD score of behavioral intention was 25.2±3.1 out of 28. In examining the difference in the mean scores of TPB constructs for smoking, those who answered “No” to the questions had a higher mean score than those who answered “Yes”, and this difference was statistically significant (P=0.01) (Table 1).

The results of the Spearman correlation test showed a significant correlation between attitude and perceived behavioral control constructs of TPB (r=0.47, P=0.01). There was a significant negative relationship between the construct of behavioral intention and the four items related to smoking (P=0.01), including experience of cigarette smoking (r=-0.348), the experience of hookah smoking (r=-0.271), current cigarette smoking (r=-0.195), and cigarette smoking in the past three months (r=-0.361) (Table 2). 

The logistic regression model of the experience cigarette smoking showed that, if the effect of other TPB constructs remained constant, the increase in the attitude (P=0.007, 95%CI; 0.85-0.97, OR=0.91) and behavioral intention (OR=0.75, 95%CI; 0.64-0.88, P=0.001) could reduce the risk of smoking cigarette in the students. The other two constructs of perceived behavioral control and subjective norms could not predict smoking in them. Logistic regression model of the experience of hookah smoking showed that, if the effect of other TPB constructs remained constant, the increase in perceived behavioral control (P=0.007, 95%CI: 0.87-0.97, OR=0.92), subjective norms (P=0.02, 95%CI: 0.70-0.97, OR=0.83), and behavioral intention (P=0.03, 95% CI: 0.78 - 0.99, OR=0.88) could reduce the risk of smoking hookah in female students. The attitude construct could not predict hookah use (Table 3).

Discussion
This study aimed to determine the predictive factors of smoking based on the TPB in female high school students. Many students had previously tried cigarette smoking, had smoked cigarettes in the past three months, and were currently smoking cigarettes and hookah. Najafi et al., in a study in high schools in Rasht, reported the prevalence of cigarette smoking for the first time as 18.3% in boys and 12.3% in girls [11]. Haghdoost et al., in a meta-analysis of Iranian university students during 2001-2011, reported the prevalence of cigarette smoking in male and female students as 19.8% and 2.2%, respectively [13]. In the study by Atai Asl et al. on high school students in Tabriz City, it was also found that 5.9% of students smoked cigarettes regularly, 15.4% had the experience of cigarette smoking, 33.5% had the experience of hookah, and 5% smoked hookah regularly [4]. Zebhi’s study on the students of Guilan University of Medical Sciences also showed that 4.9% smoked a cigarette and 30.22% smoked hookah [14]. The results of Miri Moghaddam’s study also confirmed that hookah use was higher than cigarette smoking in medical students in southern Iran and the majority of students mentioned the first experience of tobacco use around the age of 15-20 years [5]. According to these studies and the present study, tobacco use (especially hookah) is common in students in different parts of Iran (mostly start at school age). Meanwhile, in industrialized countries, the prevalence of tobacco consumption is decreasing [1]. In these countries, the replacement of traditional tobacco (cigarettes and hookahs) with other forms of substances such as electronic cigarettes [15, 16] or marijuana [17] is common.
The present study results showed that the mean scores of TPB constructs in students who did not have a history of tobacco use were significantly higher than those with a history of tobacco use. There was also a significant positive relationship between the constructs of TPB in them. In cigarette smoking, attitude and behavioral intention constructs had higher predictive power than other constructs. In hookah use, the constructs of perceived behavioral control structures, subjective norms, and behavioral intention had a higher predictive power. In other words, the TPB can predict tobacco use. Karimy et al. also found that all constructs of the TPB help predict cigarette smoking behavior, especially perceived behavioral control and attitude [10]. Topa et al. also reported that the TPB is a good predictor of tobacco use. According to them, behavioral intention and perceived behavioral control could predict smoking behavior, and the behavioral intention was a stronger one. Attitude, subjective norms, and perceived behavioral control could predict behavioral intention, where attitude had a weaker effect [18]. Joveini et al. found a significant relationship between attitude and the intention to quit or continue hookah use in the future. It means that by increasing positive attitude towards hookah use, the intention to quit decreases [19]. Su et al. also found that all constructs of TPB were significantly related to cigarette smoking behavior and intention, and a positive psychological attitude towards cigarette smoking was significantly associated with cigarette smoking behavior and intention [20]. Tapera found that the construct of TPB predicted tobacco use, and students who had a positive attitude towards cigarette smoking were more likely to smoke [8].
In the present study, perceived behavioral control and subjective norms could not predict cigarette smoking compared to the studies mentioned above. This discrepancy may be related to the difference in the number of samples that smoke. Their number was lower compared to the number of samples in above mentioned studies. Moreover, it may be because only female students participated in our study. Culturally, tobacco use by girls is perceived as a more inappropriate behavior than by boys in Iran, and subjective norms are stricter for girls. Hence, cigarette smoking in girls is less than in boys, which may influence the outcome. The reason that attitude construct could not predict hookah smoking in the present study is that the obscenity of hookah smoking in Iranian society is less than cigarette smoking, and the use of hookah as a recreational pastime is more common among adolescents and families.
Because of the increasing prevalence of tobacco use in adolescent students in Iran, and since adolescence is the age of psychological development and the beginning of many high-risk and substance-use behaviors, it is necessary to devise special programs and plans for this stage of life to investigate the factors that cause this behavior in adolescents. The TPB can be used to identify the predictors of tobacco use. For example, by adequately informing the mass media and social media about the complications of smoking, a negative attitude towards this issue can be created. Also, by teaching life skills (especially the skill of saying no) to adolescents, their perceived behavioral control can be strengthened. By creating healthy recreational environments within families and friends with easy access to leisure facilities (e.g., swimming pools, libraries, playgrounds), appropriate options for filling leisure time and discharging the energy of adolescents can be created. 
As one of the limitations of the present study, it was conducted only in public schools in Rasht, which may not be generalizable to all female adolescents. Hence, further studies are recommended in other parts of the country. Since this study was cross-sectional, analytical and interventional studies are recommended to examine the predictors of tobacco use more closely.

Ethical Considerations
Compliance with ethical guidelines
This study obtained its ethical approval from the Research Ethics Committee of Guilan University of Medical Sciences (Code: IR.GUMS.REC.2016.126).

Funding
This study was extracted from a Master’s thesis of the first author. It received funding from Guilan University of Medical Sciences. 

Authors' contributions
Conceptualization, methodology, and writing – original draft: Maryam Ghasemipour, Rabiollah Farmanbar, Parisa Kasmaee, and Sima Nickandish; Resources: Maryam Ghasemipour; Editing, review, and data analysis: Zahra Atrkar Roshan and Maryam Ghasemipour; Final approval: All authors.

Conflict of interest
The authors declared no conflict of interest.

Acknowledgements
The authors would like to thank Guilan University of Medical Sciences and the Education Organization of Rasht City for their support and cooperation.


References

1. World Health Organization (WHO). WHO global report on trends in prevalence of tobacco use 2000-2025. 3^rd ed. Geneva: World Health Organization; 2019. https://www.who.int/publications/i/item/who-global-report-on-trends-in-prevalence-of-tobacco-use-2000-2025-third-edition
2. World Health Organization (WHO). WHO report on the global tobacco epidemic, 2013: Enforcing bans on tobacco advertising, promotion and sponsorship. Geneva: World Health Organization; 2013. https://www.who.int/publications/i/item/9789241505871
3. Karimy M, Niknami S, Heidarnia AR, Hajizadeh I. [The effect of an educational package on the attitude, self efficacy and tobacco use preventive behaviors of adolescents (Persian)]. Payesh (Health Monitor). 2014; 13(4):441-8. [DOI:20.1001.1.16807626.1393.13.4.7.9]
4. Ataei Asl M, Sarbakhsh P, Dadashzadeh H, Augner C, Anbarlouei M, Mohammadpoorasl A. Relationship between happiness and tobacco smoking among high school students. Epidemiology and Health. 2018; 40:e2018009. [DOI:10.4178/epih.e2018009] [PMID] [PMCID]
5. Miri-Moghaddam M, Shahrakipour M, Nasseri S, Miri-Moghaddam E. Higher prevalence of water pipe compared to cigarette smoking among medical students in Southeast Iran. Central European Journal of Public Health. 2019; 27(3):188-94. [DOI:10.21101/cejph.a5615] [PMID]
6. Dehdari T, Jafari A, Joveyni H. [Students’ perspectives in Tehran University of Medical Sciences about factors affecting smoking hookah (Persian)]. Razi Journal of Medical Sciences. 2012; 19(95):17-24. http://rjms.iums.ac.ir/article-1-2116-fa.html
7. Ben Natan M, Golubev V, Shamrai V. Smoking during pregnancy: Analysis of influencing factors using the Theory of Planned Behaviour. International Nursing Review. 2010; 57(3):388-94. [DOI:10.1111/j.1466-7657.2010.00807.x] [PMID]
8. Tapera R, Mbongwe B, Mhaka-Mutepfa M, Lord A, Phaladze NA, Zetola NM. The theory of planned behavior as a behavior change model for tobacco control strategies among adolescents in Botswana. PloS One. 2020; 15(6):e0233462. [DOI:10.1371/journal.pone.0233462] [PMID] [PMCID]
9. Ajzen I. Behavioral intervention based on the theory of planned behavior. Theory of Planned Behavior (TPB). 2006; 1-6. https://www.researchgate.net/publication/245582784_Behavioral_Interventions_Based_on_the_Theory_of_Planned_Behavior
10. Karimy M, Zareban I, Araban M, Montazeri A. An extended Theory of Planned Behavior (TPB) used to predict smoking behavior among a sample of Iranian medical students. International Journal of High Risk Behaviors & Addiction. 2015; 4(3):e24715. [DOI:10.5812/ijhrba.24715] [PMID] [PMCID]
11. Najafi K, Avakh F, Khalkhali SMR, Nazifi F, Farahi H, Faghirpoor M. [The prevalence of substance use among high-school students in Rasht City (Iran) (Persian)]. Iranian Journal of Psychiatry and Clinical Psychology. 2005; 11(2):233-6. http://ijpcp.iums.ac.ir/article-1-77-fa.html
12. Karimy M, Niknami SH, Heidarnia AR, Hajizadeh E. [Psychometric properties of a theory of planned behavior questionnaire for tobacco use in male adolescents (Persian)]. Journal of Sabzevar University of Medical Sciences. 2012; 19(2):190-7. http://jsums.medsab.ac.ir/article_286.html?lang=fa
13. Haghdoost AA, Moosazadeh M. The prevalence of cigarette smoking among students of Iran’s universities: A systematic review and meta-analysis. Journal of Research in Medical Sciences. 2013; 18(8):717-25. [PMID] [PMCID]
14. Zebhi A, Pakseresht S, Mokhtari N, Mahdevi Roshan M, Kazemnezhad Leyli E, Monfared A. Osteoporosis preventive lifestyle and Vitamin D status among nursing students of Guilan University of Medical Sciences. Journal of Holistic Nursing and Midwifery. 2019; 29(2):113-21. [DOI:10.32598/JHNM.29.2.113]
15. Treur JL, Rozema AD, Mathijssen JJP, van Oers H, Vink JM. E-cigarette and waterpipe use in two adolescent cohorts: Cross-sectional and longitudinal associations with conventional cigarette smoking. European Journal of Epidemiology. 2018; 33(3):323-34. [DOI:10.1007/s10654-017-0345-9] [PMID] [PMCID]
16. Lee A, Lee SY, Lee KS. The use of heated tobacco products is associated with asthma, allergic rhinitis, and atopic dermatitis in Korean adolescents. Scientific Reports. 2019; 9(1):17699. [DOI:10.1038/s41598-019-54102-4] [PMID] [PMCID]
17. Dai H. Trends in single, dual, and poly use of alcohol, cigarettes, and marijuana among US high-school students: 1991-2017. American Journal of Public Health. 2019; 109(8):1138-40. [DOI:10.2105/AJPH.2019.305122] [PMID] [PMCID]
18. Topa G, Moriano JA. Theory of planned behavior and smoking: Meta-analysis and SEM model. Substance Abuse and Rehabilitation. 2010; 1:23-33. [DOI:10.2147/SAR.S15168] [PMID] [PMCID]
19. Joveyni H, Dehdari T, Gohari MR, Gharibnavaz H. [The survey of attitudes, subjective norms, and perceived behavioral control of college students about hookah smoking cessation (Persian)]. Journal of Health System Research. 2013; 8(7):1311-21. http://hsr.mui.ac.ir/files/site1/user_files_2e7b82/archive-A-10-39-384-2b6b065.pdf
20. Su X, Li L, Griffiths SM, Gao Y, Lau JTF, Mo PKH. Smoking behaviors and intentions among adolescents in rural China: The application of the Theory of Planned Behavior and the role of social influence. Addictive Behaviors. 2015; 48:44-51. [DOI:10.1016/j.addbeh.2015.04.005] [PMID]