Adolescents aged 13-17 who smoke conventional cigarettes and e-cigarettes against selected family factors, school performance and health literacy

■ Introduction

Despite the numerous educational activities and campaigns on the consequences of smoking in the United States alone, according to the Centres for Disease Control and Prevention, every day approximately 1,600 young people under the age of 18 try their first cigarette [1]. Irrespective of the risk of addiction and the adverse effects on the developing brain, smoking negatively affects learning, memory, and concentration, and stimulates openness to other types of addiction [2].
Global reports confirm the downward trend in smoking prevalence of conventional cigarettes [3, 4]. At the same time, more and more attention is being paid to substitute products, the market for which is growing rapidly and prevalence is increasing [5, 6]. A distinction is made between Electronic Nicotine Delivery Systems (ENDS) and Electronic Non- Nicotine Delivery Systems (ENNDS), of which e-cigarettes and vaporisers are the predominant forms [7]. Conclusions on the harmfulness of these products are inconclusive, although there are a growing number of reports confirming their negative health consequences [8]. In the adult population, it is not ruled out that switching to e-cigarettes may be a step towards quitting smoking. However, previous non-smokers, once they have passed the stage of using e-cigarettes, may start to smoke tobacco regularly. Research on this topic is conducted among different population groups with longitudinal studies of particular value [9]. Studies conducted among adolescents and young adults are a separate category and have a different rationale. Experts agree it is difficult to find any benefit in this age group and that the use of ENDS or ENNDS can be a gateway to many addictions [10, 11]. Additionally, attention is drawn to the lack of young people’s knowledge about the effects and harms of these products, their attractive appearance and variety of flavours, their low cost and ease of access, the possibility of discreet use in public places as well as the vulnerability of young people to advertising and the influence of peers and celebrities [12].
The Health Behaviour in School-aged Children (HBSC) international study is a unique source of information on selected health indicators, including the prevalence of psychoactive substance use. It started in the early 1980s as a project dedicated to monitoring youth smoking prevalence in Europe [13]. Over time, the geographic coverage and scope has expanded substantially, but the tobacco topic has remained the focus, although the approach to measuring smoking prevalence has changed. Questions on e-cigarette use were included in the 2017/18 HBSC survey protocol as an optional module, only to become mandatory in the next round. The multi-faceted HBSC questionnaire allows for the analysis of potential determinants of health behaviours, which are broadly divided into factors related to family, school and peer environment. The new conceptual study model, which is similar to the bio-ecological model [14], also takes into consideration a number of individual resources like academic achievement, peer bonding, spirituality or health literacy (HL) [15-17].
Regarding different population groups, much attention has been paid to the social determinants of smoking, and in the adult population it has been shown that less educated people are more likely to smoke [18]. In the case of students, educational attainment may be a measure of their target education, and in the absence of data on parental education, the mother or father’s smoking status may be a proxy socio-economic status measure. It can be hypothesised that lower cultural capital and negative family patterns are more conducive to adolescent smoking attempts than limited material resources [19, 20]. Conversely, a high level of HL, shaped by the family but also by appropriate health education at school, may be a protective factor [21]. A high level of HL may influence better perception of smoking-related harms and adoption of preventive behaviours [22].
Health literacy is a fundamental capability that ensures access to health information. Nutbeam [23] defines HL as an asset comprising a set of skills and knowledge that enables individuals to exert greater control over their health and the determinants of health. These skills and knowledge range from theoretical health understanding and practical abilities to critical thinking, self-awareness and citizenship [24].
The World Health Organization (WHO) in its guidelines on smoking entitled Framework Convention on Tobacco Control (WHO FCTC) stresses the importance of developing strategies to reduce both the demand and supply of tobacco at national, regional, and international levels [25]. The WHO indicates that actions should be aimed at protecting the youngest and thus stopping children from taking up smoking. The present study fits into these guidelines by providing detailed knowledge on the determinants of smoking in adolescents. Drawing attention to low HL as an important determinant is even more important given that the last two rounds of HBSC surveys indicate a deterioration in its levels [26].
The aim of the analyses is to determine the extent to which selected family factors and individual resources influence the prevalence of smoking among schoolchildren in Poland. The focus was on the relationship with HL and parental smoking. The following research questions were formulated:
• How does the prevalence of cigarette and e-cigarette smoking change in boys and girls seventh grade of primary school and third grade of secondary school?
• Does family wealth and social position influence the prevalence of cigarette and e-cigarette smoking among young people?
• Does parental smoking affect the prevalence of cigarette and e-cigarette smoking among young people?
• Do individual resources such as academic achievement and level of HL influence the prevalence of cigarette and e-cigarette smoking among young people?

■ Material and methods

Sample

The data are from the HBSC survey conducted in Poland in the 2017/2018 school year. A total of 5279 pupils with no missing data in key variables were included in the analyses. Three school years were included: primary school Grade VII (13-year-olds), secondary school Grade I (15-year-olds) and Grade III (17-year-olds) excluding the youngest HBSC group (11-year-olds) in which an abbreviated questionnaire was applied. The oldest group was included in the Polish study beyond the international protocol. The mean age of the subjects was 15.45 ± 1.73 years. Students qualified for the analyses presented below studied in 213 schools (333 classes) located in all provinces. Surveys were conducted in schools using the traditional Paper-and-Pencil Interviewing (PAPI) method. The characteristics of the sample are presented along with the results and there is a more detailed description of the survey procedure in the national HBSC 2018 report [27].
In addition, HBSC data were compared with ESPAD (European School Survey Project on Alcohol and other Drugs) data collected in 2019 in a nationwide sample of 91 secondary school second graders (N = 2372) based on published data [28].

Tools, variables, and indicators

Students were asked on how many days they had smoked cigarettes and e-cigarettes. They answered separately from a lifetime perspective and the last 30 days, with seven response categories: never, 1-2 days, 3-5 days, 6-9 days, 10-19 days, 20-29 days and 30 days or more. These were questions adapted to the HBSC protocol from the ESPAD survey but were not identical [28, 29] with the number of days rather than the number of times considered and the question layout was the same for both types of cigarettes. After recoding, non-smokers, occasional smokers (1-2 days) and more frequent smokers were distinguished. Based on the original layout of responses, indices of smoking the two types of cigarettes in the lifetime (INDLT) and in the last 30 days (IND30) were constructed using the PCA (principal component analysis) method. The higher the value, the more frequent the declared smoking of tobacco or electronic cigarettes. Gender, age (coded into three groups) and place of residence (divided into urban and rural areas) were considered as basic demographic characteristics. When determining the family situation, the family’s material resources and subjectively assessed social position were considered. The Family Affluence Scale (FAS) in its current third version consists of six questions relating to: number of vehicles, bedroom sharing, computer ownership, bathrooms at home, dishwashers at home, and family vacations. The FAS takes a range of 0-13 points and is conventionally divided into three ranges [30]. The subjective social position (SSP) of the family was determined by the students on a visual scale taking the form of a ladder with a range of values from 0 to 10. SSP correlates with FAS at ρ = 0.326, which mandates an independent study of the impact of these two factors on the variation in smoking indices. The student’s individual resources were determined by the level of HL and subjective assessment of academic achievement. A ten-item tool implemented by the HBSC research network, known as HLSAC-10 (Health Literacy in School-aged Children), is used to measure HL. It contains two questions each from five dimensions of HL: theoretical knowledge, practical knowledge, critical thinking, self-awareness and citizenship. The overall index has good psychometric properties, with α Cronbach’s coefficient of 0.839 in the group analysed. The summative scale takes a range of 10-40 points and is conventionally divided into three ranges [31]. School achievement was measured on a visual scale, which was previously validated in Poland with the results of state tests [32].
The teenagers surveyed were also asked: does your Mum or Dad smoke cigarettes? – with response categories: does not smoke, sometimes, daily, don’t know, don’t have or don’t see this person. The last two categories were combined as unknown smoking status. Data on each parent were analysed separately.
As additional Polish questions, place of residence, family social position, academic achievement and parental smoking, were included from beyond the HBSC 2017/18 international protocol.

Statistical analysis

The distribution of the categorised variables in the study sample and the mean values of the continuous indices with standard deviation (SD) are presented. The relationship between categorised variables was tested with the χ² test for independent or dependent data (McNemar test). Correlations between continuous variables were tested with the Spearman ρ coefficient. Mean INDLT and IND30 smoking indices were compared between groups distinguished by underlying factors. These averages were estimated as theoretical values in a general linear model (GLM), including gender, age, residence, family wealth according to FAS and its social position, school achievements, HL and parental smoking as independent variables. Mean values were given with 95% confidence intervals and differences between groups were analysed with an F-test (ANOVA) showing the significance of the main effects of the GLM model. Differences between pairs of groups were tested with a multiple comparisons test analysis. The goodness-of-fit of the models was assessed with the R² coefficient. Calculations were performed using the SPSS v.29 package (for GLM, UNIANOVA procedure).

■ Results

Distribution of underlying factors

The characteristics of the sample can be found in Table I. There were 46.7% boys and 53.3% girls. The sample size varied for three grades, with the strongest representation of pupils in primary VII classes. Residents of urban areas made up 60.5% of the sample, which is representative of the country. The mean of the FASIII wealth scale was 7.80 ± 2.30 and, according to the criteria used in the HBSC research network studies, almost one in three students grew up in a relatively poor family. The teenagers rated the social position of their family at an average of 7.15 points (SD = 1.82), and one in five respondents scored lower than 6 on this scale. A similar proportion of students felt they had poor academic performance, with an average score of 6.23 (SD = 2.06) on the school achievement scale. HL was rated at an average of 30.69 points (SD = 4.51), with 10.8% of respondents categorised as low HL.

Smoking in the family

The smoking intensity indices significantly (p < 0.001) and positively correlate with age (ρ equal to 0.324 for INDLT and 0.293 for IND30 respectively) and negatively with educational attainment (–0.210 and –0.197). There was a weak positive correlation of INDLT with family wealth (ρ = 0.030; p = 0.027) and no significant association of FAS with IND30 (p = 0.324). There was a significant although weak negative correlation with family social position (–0.077 for INDLT and –0.052 for IND30). The association with HL is also negative but even weaker (ρ equal to –0.031 for INDLT and –0.044 for IND30 respectively). More interestingly, it is significant in girls, but no longer significant in boys. Many other correlations are stronger in girls. For example, the correlation of academic achievement with HL is 0.225 compared to 0.189 for boys.
The data on adolescent smoking confirm that all indicators change significantly with age (Table II). The percentage of adolescents who had smoked conventional cigarettes ever in their lives was 17.2%, 36.3% and 52.9% in the following age groups. For the last 30 days, the corresponding frequencies were 6.4%, 16.8% and 32.5%. No significant gender-related differences were found at the extremes of the age groups. Only among 15-year-olds were girls more likely than boys to have smoked conventional cigarettes in the last 30 days (14.5% vs. 18.5%, p = 0.034) with no differences in INDLT.
percentage of young people who had ever used e-cigarettes in their lives was 21.8%, 41.5% and 49.7% in the following age groups. 8.7%, 18.8% and 28.6% of respondents admitted to having used them in the last 30 days respectively. In the case of electronic cigarettes, gender differences persist, but only among 13-year-olds and 17-year-olds, where boys are more likely to use them than girls. In the 15-year-old group, gender-related differences level off (for lifetime experience p = 0.273, and for the last 30 days p = 0.508). Thus, in the two younger age groups, e-cigarettes are more popular than conventional cigarettes while in 17-year-olds, the prevalence of smoking conventional cigarettes is higher.
The prevalence of smoking of the two types of cigarettes was compared across age groups, separately for both sexes and the two-time perspectives, using McNemar’s test for linked data and using three categories as in Table II. Among boys, significant (p < 0.001) differences persisted in the two younger age groups with a prevalence of e-cigarette smoking. In 17-year-old boys, the differences waned, with a p = 0.090 for lifetime rates and p = 0.097 for the last 30 days when tested. In girls, on the contrary, significant differences (p < 0.001) only appear in 17-year-old girls with a prevalence of smoking conventional cigarettes.
Table III shows the parental smoking prevalence as determined from the child’s perspective. Regular or occasional smoking was reported by 23.6% of mothers and 31.6% of fathers, including daily smoking by 16.1% and 23.2% respectively. Adolescents had more difficulty identifying the smoking status of fathers than mothers though overall data gaps were small. Combining information on both parents, 57.6% of adolescents were found to have both non-smoking parents (or the only parent they have contact with), missing data on both parents was reported by only 0.6% and the remaining 41.8% had at least one smoking parent. It was also found that the prevalence of maternal and paternal smoking varied significantly (p < 0.001) according to family wealth and social position (Table III). The percentage of smoking parents clearly decreases in the wealthiest families and when the social position of the family was defined as high. Place of residence is a differentiating factor for mothers’ smoking prevalence (p = 0.006) with no significant differences for fathers (p = 0.675).
It is also noteworthy that there was a significant association between child’s HL and the frequency of parental smoking. The percentage with low HL level decreased to 9.9% and 10.1% in case of non-smoking mother and father respectively.

A comparison of HBSC and ESPAD data related to smoking and vaping

An interesting extension to the analysis of age-related changes in smoking prevalence of traditional cigarettes and e-cigarettes is the inclusion of 16-year-olds surveyed within the ESPAD study in 2019. However, it is worth recalling here the previously described differences in the way the questions were worded. A clear age-related trend persists for both indicators on traditional cigarettes smoking. There is no difference between 16- and 17-year-olds for current e-cigarettes use. Over their lifetime, 16-year-olds (ESPAD) reported more frequent vaping than 17-year-olds (HBSC). It’s worth noting that the students surveyed in the 2019 ESPAD survey came from the same cohort as the 15-year-olds participating a year earlier in the HBSC survey (Figure 1).

Factors influencing the variability of smoking indices

There is a strong association between the frequency of smoking conventional cigarettes and e-cigarettes. In the study group, 56.2% of respondents had smoked neither in their lifetime, 27.4% had used both types of cigarettes and the remaining 16.4% had used only one type. In contrast, 74.8% of respondents had smoked neither conventional cigarettes nor e-cigarettes in the past 30 days, 10.5% both types and 14.7% only one of the types. Standardised INDLT and IND30 indices, which also consider the frequency of smoking, were used to examine the combined burden of these two behaviours on young people and its determinants. They are correlated with each other at ρ = 0.754.
Table IV shows comparisons between different social groups in terms of the mean INDLT indicators relating to lifetime smoking. It is worth noting that the mean values are adjusted for all factors under study. After adjusting for other factors, the overall mean INDLT is equal to 0.193 (95% CI: 0.126-0.260). The GLM model explains 14.7% of the variability of this index. A significant relationship was found with age to the detriment of older adolescents and with academic achievement to the detriment of poorer students. Post-hoc analysis confirmed that in both cases, the means were significantly different across the three groups of students (p < 0.001). The association with family wealth was found to be significant, with adolescents from more affluent families at greater risk of smoking. The first wealth group is significantly (p < 0.001) different from the other two in terms of INDLT levels, and the difference between the average and wealthiest group is also significant (p = 0.036). No significant relationship with the gender of the respondents and place of residence was confirmed. In the case of family social position and adolescent HL, no statistically significant relationship was obtained either. However, a significant relationship with parental smoking was confirmed. INDLT takes on the most favourable, as the lowest values in the group of adolescents whose parents do not smoke. Particularly burdening is the daily smoking of the mother and father. Adolescents who were unable to determine their mother’s smoking status are also more likely to smoke, while the effect of missing data on this subject appears weaker for fathers. In a multiple comparison analysis, adolescents with a non-smoking mother achieved significantly lower (p < 0.001) INDLT values than peers whose mothers smoke daily or when the mother’s smoking status was unknown. No difference was shown when comparing adolescents with non-smoking mothers and with mothers smoking occasionally (p = 0.145). The difference in terms of INDLT between those having mothers who smoke occasionally and daily persisted (p = 0.008). In contrast, adolescents with non-smoking fathers obtained lower INDLT values than those with occasionally or daily smoking fathers. No difference was shown with respect to the group with unspecified father smoking status (p = 0.198).
Table V similarly compares the average IND30 indices relating to the last 30 days. These are also the values adjusted for all factors analysed. After adjusting for other factors, the overall mean IND30 is 0.259 (95% CI: 0.190-0.328). This model determines 10.5% of the variability of the index under study. Inferences regarding IND30 determinants change significantly compared to INDLT determinants, with only place of residence remaining a non-significant factor. This was less favourable, as higher IND30 values are achieved by boys compared to girls and by older compared to younger adolescents. The prevalence of current smoking increases as the family’s financial situation improves, but also as academic performance deteriorates. However, the difference between young people from the most affluent and average affluent families is weaker than for INDLT (p = 0.056). The association with family social position is significant but non-linear, with young people from families rated as average being in the most favourable condition. Only the difference between the group with the best and average social position was found to be significant in the multiple comparisons test (p = 0.006). In the case of HL, low HL is an aggravating factor, with no difference in the current smoking prevalence of adolescents with average and high HL (p = 0.638). Children whose mothers and fathers do not smoke scored large lower mean IND30 values, compared to peers whose parents smoke daily. It is also worth noting the very high value of this index in the small group of adolescents who could not determine whether their mother was a smoker. Adolescents with non-smoking mothers achieved significantly lower IND30 values than those with mothers who smoke daily or in the absence of data on maternal smoking (p < 0.001). The second group, those having mothers who smoke occasionally, differed only from the fourth group (p = 0.005). Adolescents with non-smoking fathers achieved significantly lower IND30 values than those with fathers who smoke daily or occasionally, with no differences from group four (p = 0.294). In the absence of father smoking data, there were no significant differences in IND30 levels compared to the other three groups.

■ Discussion

The study presented herein included 5279 Polish students aged 13-17 participating in the 2017/18 HBSC international survey. The article authors were particularly interested in the relationship between adolescents’ levels of health literacy (HL) and their parents’ smoking and frequency of use of conventional cigarettes and e-cigarettes. According to ESPAD 2019 data, adolescent smoking is more common in Poland than in many European countries. Among 16-year-olds, Polish adolescents ranked fourth out of 35 countries for decreasing smoking rates [29]. It is difficult to compare the frequencies we obtained with the ESPAD study due to differences in the way the questions were asked especially about e-cigarettes.
In answering the first research question on the demographic determinants of cigarette smoking, the authors showed that conventional cigarettes are more likely to be used by older adolescents, while younger adolescents are more likely to use e-cigarettes. It can be hypothesised that e-cigarettes and vaporisers are more popular with younger adolescents, as their arrival on the market occurred more recently when older adolescents were already past tobacco initiation [33]. In addition, it is not insignificant that smoking e-cigarettes and vaping are not associated with the noxious smell and taste of burnt tobacco, but rather can be associated with pleasant flavours. E-cigarettes and vaporisers are available in flavoured versions similar to the smell and taste of fruit or even candy, which may encourage children. E-cigarettes are surging in popularity among teenagers for many reasons, including their availability, variety of flavours and possibility of discreet use beyond parental control. Moreover, expose to e-cigarette advertising may change young people’s harm perceptions [34]. This is particularly dangerous as vaping is a prelude to smoking conventional cigarettes and trying other stimulants as revealed in a Canadian study of a 2019 sample of 13,602 teenagers aged 12-17 years [35].
When comparing smoking prevalence in both sexes, the study authors highlighted the middle-age group. Only among 15-year-olds were girls more likely than boys to have smoked conventional cigarettes in the last 30 days, and they also did not differ from boys in terms of their lifetime smoking frequency. The group of adolescent girls is increasingly described as the one more likely to engage in risky behaviour [36]. It is worth advising in-depth research and qualitative studies to better understand the reasons behind this.
When looking for an answer to the second question concerning the relationship with family socio-economic status, it was confirmed that adolescents from wealthier families smoke more often. This is consistent with the results of earlier international analyses based on HBSC data, where it was additionally found that the gap between adolescents growing up in poorer and richer families increases in more economically developed countries [37]. The fact that more affluent adolescents in Poland smoke cigarettes shows the importance of the state regulating tobacco availability. The high price of cigarettes or e-cigarettes effectively discourages a part of the population. Other researchers emphasised that the economic status of adolescents matters, but not as much as among adult smokers, where the problem increased in less affluent families [38]. In the light of our results, the inference about the influence of family social position is different from that of wealth. The prevalence of adolescent smoking increases in the lowest and highest status families, and there may be different determinants in each of these groups.
The third research question concerned parental smoking and its impact on adolescent smoking of cigarettes and e-cigarettes. Lower INDLT and IND30 indices in the group of adolescents with non-smoking parents were confirmed. Fleary et al. [39], also showed an association between parents’ and children’s smoking behaviour on the base of observation of 300 child-parent dyads. The results obtained in the present study indicate a strong social gradient in parental smoking inverse to that shown for the determinants of adolescent smoking. It is still important to run educational campaigns targeting different groups in society including smoking parents. The communication message should address not only the clinical consequences of their children’s passive smoking, but also the risk of their children falling into the same unhealthy habit. Attention was also drawn to the higher risk of smoking among adolescents with poor contact with their parents or unable to determine the smoking status of their father and mother. The higher prevalence of smoking in single-parent families is confirmed by the results of other, some also Polish, studies based on earlier editions of the HBSC [40].
Referring to the fourth research question, a significant relationship between adolescent smoking frequency and health competences was confirmed, but only in relation to current smoking. In turn, academic achievement influenced the variability of both indices. Attention was also drawn to the interrelation of these two variables and their relationship with family situation. Thierry Gagne et al. [20] suggest that a new measure of cultural capital, which combines health values, education and knowledge and family resources, should be introduced into research on the social determinants of risk behaviour. Our study, by introducing family characteristics and individual resources of young people, particularly their level of HL is consistent with this approach. The assessment of the impact of HL level on risk behaviour taking is a continuation of previous analyses conducted by the Polish HBSC team also based on the HBSC 2017/2018 data. The previous analyses investigated the determinants of the risk behaviour syndrome in 17-year-olds including, in addition to smoking and e-cigarettes, drinking alcohol, binge drinking and marijuana use. After adjusting for other factors, HL level was shown to remain a significant predictor of the severity of this syndrome [41]. Similar conclusions were reached by researchers from Portugal, who examined functional HL using NVS-PTeen instrument (Newest Vital Sign for Portuguese Adolescents). They confirmed that the education of the adolescents measured by the number of grades completed an HL academic performance (mathematics and Portuguese language) are related to the level and shape selected health outcomes [42]. The strong association of poorer academic achievement with adolescent smoking confirmed in our study is also indicated by the results of a survey of adolescents aged 14-16 years in six countries as part of the SILNE (Smoking Inequalities: Learning from Natural Experiments) project, which used the results of the 2013/14 HBSC survey as an additional data source. Not only was a protective effect of school achievement against smoking prevalence proven, but also against the risk of nicotine dependence [43].
Limitation. The analyses conducted have several strengths, but limitations should also be noted. The data were collected in 2018, so the current smoking prevalence of conventional cigarettes and e-cigarettes by young people, as well as by parents, may be different. Newer generations of products replacing conventional cigarettes are entering the market [12]. However, the results of the next round of 2022 are yet to be published, and the frequencies obtained in these more recent studies are burdened by the fact that the data were collected during the COVID-19 pandemic. The period of data collection is not crucial regarding analyses of associations between variables, and some analyses could not be repeated on more recent data due to the revised scope of the Polish questionnaire. The cross-sectional nature of the HBSC surveys is also a limitation of the above analyses, which, in the absence of an appropriate time sequence, limits causal inference. When examining determinants, a time perspective (lifetime vs. current state) was used as a basis, without delving into the differences in determinants of electronic and conventional cigarette smoking. Another limitation was the generic nature of the HLSAC measurement tool without reference to smoking.
It is also worth noting the analytical approach used in combining information on dual use, which can be considered a strength of the analyses. In this study, standardised indices measuring not only the fact of smoking, but also its frequency, were used. Similar indices have been used in other studies on risk behaviour as a method of information reduction to facilitate inference about determinants [41]. In addition, the values of these indices were adjusted for the factors analysed, allowing a more reliable comparison of mean values, unencumbered by the distribution of other factors.

■ Implications for further research and practice

As health literacy (HL) was one of the leading themes of the analysis and the inspiration for this topic, it is worth referring in the conclusions to the definition of HL, which indicates that it is people’s knowledge, motivation, and competence to access, understand, evaluate, and apply health information resulting in appropriate health decisions in everyday life. According to this definition, educational activities to raise health awareness among young people regarding cigarette smoking and the use of newer generations of substitute products should be based on:
Universal access to knowledge for students, parents, teachers, and the community, as a so-called whole-school-approach [44]. Education should ensure the development of a critical understanding of the information provided through advertising or media messages. Educational efforts to reduce youth smoking should motivate them to make healthy choices and adopt healthy lifestyles. This is one of the most difficult challenges as creating trends among young people also requires the commitment of significant financial resources for health-promoting social media campaigns.
The multifaceted determinants of smoking in relation to the reasons for starting and continuing to smoke cigarettes and e-cigarettes should be kept in mind when developing educational programmes or information strategies on smoking. The data obtained in this study indicate the need for actions targeting social groups where smoking prevalence is higher.
The role of parents, demonstrated in this work, should be used as part of efforts to raise awareness of how their adverse behaviours can shape the competences and model health behaviours of children, projecting current and future health.

Conflict of interest/Konflikt interesów

None declared./Nie występuje.

Financial support/Finansowanie

None declared./Nie zadeklarowano.

Ethics/Etyka

The Bioethics Committee at the Institute of Mother and Child gave consent to conduct the research discussed, document No. 17/2017 with an annex of March 30, 2017.
Na przeprowadzenie badania zgodę wyraziła Komisja Bioetyczna przy Instytucie Matki i Dziecka, dok. nr 17/2017 z aneksem z 30.03.2017.

The work described in this article has been carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) on medical research involving human subjects, Uniform Requirements for manuscripts submitted to biomedical journals and the ethical principles defined in the Farmington Consensus of 1997.
Treści przedstawione w pracy są zgodne z zasadami Deklaracji Helsińskiej odnoszącymi się do badań z udziałem ludzi, ujednoliconymi wymaganiami dla czasopism biomedycznych oraz z zasadami etycznymi określonymi w Porozumieniu z Farmington w 1997 roku.

References/Piśmiennictwo