Susceptibility to smoking and associated factors among the youth in central and eastern European countries.

Tobacco use among young people still remains a major public health problem. The aim of this study was to examine the association between a variety of factors and susceptibility to smoking initiation and experimentation among the youth from central and eastern European countries. The data used in the current analysis, focusing on current non-smokers, is available from the Global Youth Tobacco Survey, which was performed in ve countries (the Czech Republic (2016), n = 3191; Slovakia (2016), n = 3178; Slovenia (2017), n = 2255; Romania (2017), n = 4681; Lithuania (2018), n = 2260).

of smoking in the future [8]. Knowing them is crucial for development of effective preventive measures to reduce smoking among young people, and thus, reduce the frequency of smoking.
According to its de nition, susceptibility to smoking is the lack of a rm commitment not to smoke and it is a strong predictor of regular smoking and addiction [13]. Studies indicate that the youth who are susceptible to smoking are more likely to experiment with tobacco and to become regular smokers compared to those non-susceptible ones. Moreover, susceptibility to smoking has been shown to be modi able through interventions (which can either prevent the youth from becoming susceptible to smoking or prevent susceptible adolescents from progressing to regular smokers) [14]. Various factors are indicated as correlates of smoking susceptibility among the youth, including: individual characteristics (e.g. age, gender), social environment and social contexts (e.g. family, friends, school) [8, [15][16][17][18][19][20][21]. However, such factors can be region or country speci c. GYTS, which has been conducted recently in central and eastern European countries, create a unique opportunity to evaluate or update the factors that predispose young people to be susceptible to smoking, which in turn can be useful for development and implementation of effective tobacco control strategies in these countries [12].
Thus, the aim of the study was to assess the association between a variety of factors and smoking susceptibility in the non-currently smoking youth from the Czech Republic, Slovakia, Lithuania, Slovenia and Romania.

Study design and population
The data used in the current analysis is from GYTS, a part of Global Tobacco Surveillance System (GTSS), which was developed to track tobacco use among young people and enhance the capacity of countries to design, implement and evaluate tobacco control and prevention programmes [12,18]. GYTS is a cross-sectional, nationally representative school-based survey that collects data using standardized methodology for constructing the sample frame, selecting schools and classes and processing data. Following GYTS twostage sample design, in all of the ve countries (the Czech Republic (2016), Slovakia (2016), Slovenia (2017), Romania (2017), Lithuania (2018)) schools were selected with a probability proportional to the enrollment size and within these schools classes were randomly chosen. All of the students from the selected classes were invited to participate in the survey. The overall response rates were, as follows: the Czech Republic 78.3%, Slovakia 81.7%, Slovenia 68.0%, Romania 88.6%, Lithuania 82.7%. Taking into account the study purpose, so as to assess patterns of smoking susceptibility, the analysis was restricted to the current non-smokers understood as the students who did not smoke cigarettes in the past 30 days. The participants were also excluded in case of missing data for the smoking status or a dependent variable. That resulted in the following sample size considered in the current analysis: 3191 of 3926 students from the Czech Republic, 3178 of 3997 students from Slovakia, 2255 of 2629 students from Slovenia, 4681 of 5409 students from Romania and 2260 of 3030 students from Lithuania.
The Ministry of Health of each country (Public Health Authority of the Slovak Republic) handled scienti c, ethical and technical coordination of the study. Ethical approvals were obtained from the relevant committees.
Questionnaire, dependent and independent variables GYTS uses a standard, anonymous, self-administered questionnaire that produces estimates on tobacco use (smoking and smokeless), exposure to second hand smoke, tobacco use cessation, access and availability to tobacco, awareness of anti-tobacco information and exposure to tobacco marketing [18,22]. The questionnaires and databases are publicly available at https://www.cdc.gov/tobacco/global/gtss/gtssdata/index.html [12]. The dependent variable considered in the current study was susceptibility to tobacco product use as adopted from   [13]. It is based on two questions: "If one of your best friends offered you a tobacco product, would you use it? (in Slovenia the following question was asked: "If one of your best friends offered you a cigarette or another tobacco product, would you use it? ") and "At any time during the next 12 months do you think you will use any form of tobacco?" (in Slovenia the following question was asked: "At any time during the next 12 months do you think you will smoke tobacco?"). For each of the questions the following four answers were possible: "De nitely not", "Probably not", "Probably yes" and "De nitely yes". The students who answered de nitely not to both questions were coded as non-susceptible ones and all the other youth were considered as susceptible to smoking [7]. Table S1 presents independent variables together with the codes for original questions and the analysis variables. They include sociodemographic and economic data: sex, age, parental education, money available for own expenses; information related to second hand smoking; knowledge and attitudes regarding tobacco use, pro and anti-tobacco media and advertising, and antismoking education provided by the school or family. Statistical analysis STATISTICA version 10.0 (Dell Software, Arizona, CA, USA) was used to perform the statistical analysis. Similarly to the previously published studies in this eld, the statistical analyses covered several steps [7,18]. Initially, a descriptive analysis for all the variables involved in the analysis was completed. The univariable and multivariable logistic regression analyses with results being presented as odds ratios (OR) with 95% con dence intervals were applied to study the factors linked to susceptibility to smoking among the youth in the ve central and eastern European countries. In the multivariable analyses, all the variables (p < 0.1) were simultaneously included. To test multicollinearity between the variables, the variance in ation factor (VIF) was calculated. A separate analysis was performed among the never (those who had never tried smoking in their lifetime) and ever smokers (those who had ever smoked even once one or two puffs in the past but are not current smokers). In that perspective susceptibility was used as a predictor for experimentation with tobacco products (the ever smokers) and smoking initiation (the never smokers) [7]. A p value of less than 0.05 was considered statistically signi cant.

Results
Characteristics of the students included in the current analysis are summarized in Table S2 -never smokers and Table S3 -ever smokers. Among the current non-smokers 59% (in Lithuania) to 79% (in Romania) of the students indicated the never-smoking status (p < 0.001).
The rest experimented with tobacco smoking (ever smokers) ( Table 1). Signi cant differences (p < 0.001) were observed between the countries with respect to the percentages of young people that were susceptible to smoking, from 23% in Romania to 38% in the Czech Republic (no differences were noted only between the Czech Republic and Slovakia, the Factors associated with susceptibility to tobacco products use among the never smokers Among the never smokers close to one fourth were susceptible to smoking in 4 out of 5 countries (Table S2). A lower percentage of the students classi ed as susceptible to smoking was observed in Romania (16%). Results from the univariable analysis that was run to identify correlates of smoking susceptibility are presented in Table S4. The further analysis, using the multiple regression, provided the results that are consistent among all the countries (

SHS -secondhand smoke
Factors associated with susceptibility to tobacco products use among the ever smokers About 60% of the students in the Czech Republic, Slovakia and Slovenia and about 50% of the students in Lithuania and Romania were susceptible to smoking experimentation (the analysis among the ever smokers) (Table S3). Generally, the factors predisposing to smoking experimentation were similar to those important for smoking initiation in the univariable (Table S4) and multivariable ( 2) were at risk of smoking experimentation. All the students, except for Romanian ones, who shared an opinion that smoking helped people feel more comfortable at celebrations, parties or in other social gatherings were at higher risk of smoking susceptibility. Finally, the youth from Lithuania and Romania who thought that smoking tobacco made young people look less attractive had a lower risk of smoking experimentation (Lithuania: OR = 0.7; 95%CI 0.5-1.0, Romania: OR = 0.5; 95%CI 0.4-0.7).

Discussion
Our results indicate that nearly a quarter of the students were susceptible to smoking initiation in 4 out of 5 analyzed countries (with slightly lower percentages observed in Romania). A substantial proportion of the adolescents were also susceptible to smoking experimentation (from 49-61%). The correlates of susceptibility to smoking are consistent across all the analyzed countries. The current results present target groups, such as girls, and priorities, such as strengthening knowledge and attitudes towards smoking, for preventive measures to be taken to decrease percentage of the youth who initiate or experiment with smoking.
The earlier analysis, based on GYTS in 168 countries, has indicated that around 1 in 8 never-smoking young people worldwide is susceptible to smoking, with the highest percentage of the youth declaring smoking susceptibility reported in America and Europe [18]. This is proven by the WHO estimates indicating that smoking prevalence among adolescents and adults in Europe is among the highest in WHO regions [23]. Another assessment based on GYTS data from 25 European countries has also con rmed high susceptibility to smoking among adolescents [24]. Moreover, the study conducted in a rural area of Poland (between 2014 and 2015) based on GYTS questionnaire indicated 22% of never smokers and 57% of ever smokers susceptible to smoking, which is similar to the results obtained within the current assessments [7]. Some differences between the studies concerning proportion of those susceptible to smoking can result from several reasons, among which social and cultural norms, tobacco industry in uence and legislation as well as preventive activities are the most frequently pointed ones. Moreover, age of a youth, de nition of a current non-smoking status, susceptibility to smoking (focusing only on cigarettes or on any tobacco products; types of questions used for assessment) and time period of the study are among other factors.
Consistently, in all the analyzed countries, girls were more susceptible to smoking initiation and experimentation than boys. In some studies, similar results have been observed [25][26][27][28][29], whereas in others, including that conducted in Poland among a rural population, boys have been more susceptible to smoking [7,[30][31][32][33]. Earlier analyses of GYTS data from European countries (including all 5 countries involved in the current analysis) have also con rmed a similar pattern of smoking susceptibility among the never-smoking youth to the one observed in the current assessments [24]. The differences between the countries in the obtained results can be explained by general differences between the countries and the time when the studies were conducted, which in turn, can show target groups for tobacco industry (which are currently girls) and how the youth are susceptible to tobacco marketing [30].
Pocket money turned out to be a risk factor for smoking susceptibility in our study. This nding is in line with GYTS results from Greece, Cyprus or Bangladesh [34][35][36][37]. Parents should pay more attention to how their children's pocket money is used. Furthermore, the role of this factor in susceptibility to smoking needs to be considered together with teenagers' easy access to tobacco products.
We found that adolescents whose siblings or closest friends are smokers were at higher risk of smoking initiation or experimentation comparing to those with non-smoking peers. In the study conducted in Poland, friends' smoking status was a stronger predictor of susceptibility to smoking than the parental smoking status [7,18]. Other studies also indicate that the strongest determinant of current smoking is peer in uence [34,35,[38][39][40][41][42]. Studies indicate that people tend to choose their friends based on shared characteristics, including tobacco smoking. However, having close friends who smoke does not need to mean that they cause a person to smoke. On the other hand, strong commitment not to smoke if offered a cigarette by a friend is crucial as a protective factor for not starting smoking [7]. There is also a need for peer education to help them quit smoking or at least abstinent from smoking in front of non-smokers. Creating a non-smoking fashion among young people seems to be most desirable. This has been also proven by our results that indicate that those who declared that tobacco helped people feel more comfortable at celebrations, parties or in other social gatherings were at higher risk of smoking susceptibility. Moreover, the youth who thought that people who smoke had fewer friends and that smoking made the youth less attractive had a lower risk of smoking susceptibility. So efforts should focus on changing such perceptions, norms and acceptance of smoking.
Exposure to passive smoking in public places was associated with increased susceptibility to smoking among the studied population, which is in agreement with other studies in this eld and supports the need to create smoke-free environments worldwide together with effective enforcement of existing legislation [18].
Lack of knowledge about harmful effects of passive smoking and lack of training in the eld of health consequences of tobacco smoking in the school curriculum is an additional, important factor of susceptibility to cigarette smoking. Similar results have been observed in other assessments in this eld [7,43]. Imparting knowledge about harmfulness of smoking has been one of the key tobacco control strategies.
Despite the existing legislation, our results, being in agreement with similar studies in this eld, indicate that presence of smoking in movies is associated with an increased risk of smoking initiation, experimentation and tobacco consumption in adolescents [44][45][46]. Tobacco use by celebrities in the mass media should be reduced as media play an important role in shaping personality, especially in adolescents and young adults [47]. It has been also shown that susceptibility to smoking is in uenced by exposure to advertising at points of sale [48][49][50]. Advertising at points of sale still remains a challenge. The results of research suggest that the complete ban on tobacco advertising in points of sale is signi cantly related to the reduced experimental smoking among adolescents and that this relationship is visible for both sexes [51]. Tobacco industry promotion efforts can be prevented by developing and implementing comprehensive bans along with stronger regulations, including enforcement of law, of tobacco company practices in accordance with WHO Framework Convention on Tobacco Control (FCTC) [52].
The current analysis has several strengths. GYTS is a cross-sectional, nationally representative survey and covers a large number of respondents from an adolescent population, assuring reliability and validity of the results. The protocols and questionnaires in all the analyzed countries were based on GYTS standards developed by experts in the eld, which enables a direct comparison between the countries and trends assessments. Moreover, susceptibility to smoking was assessed by two questions which constitute frequently used and reliable measure of predisposition to smoking initiation and experimentation. Finally, the analysis considers a number of various potential correlates of susceptibility to smoking including socio-demographic, economic factors, information related to second hand smoking, knowledge and attitudes regarding tobacco use, pro and anti-tobacco media, advertising and educational issues related to health consequences of smoking. Limitations of the study also need to be pointed out. Firstly, due to the cross-sectional nature of the study, claims of causation cannot be made about the observed relationships between susceptibility to smoking and the studied variables. Secondly, all the estimates in our assessment were based on self-reports, which might be affected by reporting bias. Moreover, some differences in the response rates exists between the countries (from 68% in Slovenia to 89% in Romania), which may bias the obtained results. The willingness to participate in the study can be determined by each country's socio-cultural norms, the level of trust and acceptance of being interviewed on sensitive issues. GYTS questionnaire contains some core questions. Additional questions can be selected or added by the experts depending on the country or speci c issues studied. That is why some of the variables were not available in all the analyzed populations. Finally, our analysis did not control for other substances use such as alcohol or illicit drugs, which are also indicated to be associated with smoking. Despite the limitations stated above, this study provides an important insight into the prevalence and factors associated with susceptibility to smoking in central and eastern European countries. These correlates are crucial for effectiveness of prevention strategies to be taken among the youth.

Conclusions
Nearly a quarter of the students were susceptive to smoking initiation in 4 out of 5 analyzed countries (16% of susceptible to cigarette smoking were identi ed in Romania). A substantial proportion of adolescents was also susceptible to smoking experimentation. The results obtained within this analysis are consistent across all the analyzed countries and they indicate target groups, such as girls, and priorities, such as strengthening knowledge and attitudes towards smoking, for interventional activities aimed at decreasing percentage of the youth who initiate or experiment with smoking. Declarations