Article Text
PDF
Abstract
Objective To review the available evidence for home smoking restrictions as a useful tool in the prevention of youth smoking and to make recommendations for further research.
Methods A PubMed search (1 January 1990 to 26 January 2010) identified studies involving youth ≤18 years using extensive criteria. In all, 17 studies relating home smoking restrictions to youth smoking behaviour were identified from titles, abstracts or the full text, as required. Two additional articles were identified by other means. KE, NS and EG reviewed the studies. Differences in interpretation were resolved by discussion, with EG making final decisions.
Results Of the 19 studies, 16 (including the only 2 longitudinal studies) showed at least marginal evidence of an association of home smoking restrictions with reduced adolescent smoking behaviours. Associations were more numerous and stronger in homes without adult smokers, suggesting that even in such homes, lack of a smoke-free home may undermine the parental value of not smoking. Definitions of home smoking rules, adolescent smoking behaviour and treatment of parental smoking varied widely among studies. It is recommend that future research: (1) contrast smoke-free homes for everyone against all others, (2) included an interaction term for parental smoking and having a smoke-free home, or conduct separate analyses for homes with and without parental or other adults smokers and (3) examine early and later stages of the smoking uptake continuum.
Conclusions While the evidence is suggestive for an effect, further research is required to establish causality using longitudinal designs.
- Environment
Statistics from Altmetric.com
Introduction
Youth smoking initiation is an important public health concern worldwide, considering tobacco use is the leading cause of preventable death in the world today.1 While cigarette smoking is highly addictive, individuals who have not initiated smoking by age 21 years are unlikely to ever begin. Further, the younger the age when people initiate, the more likely they will become regular smokers and the less likely it is they will ever quit.1 In the USA, Healthy People 2010 goals include reducing youth smoking by over half (from 35% to 16%).2
The first paper hypothesising that smoking restrictions in the workplace and on college campuses might reduce youth smoking was published in 1991.3 Since then, there has been growing and consistent evidence that workplace and home smoking restrictions protect non-smokers from secondhand smoke and influence the smoking behaviour of adult smokers.4 5 Our review summarises the published evidence to date on the association of smoke-free homes on youth smoking behaviours, particularly initiation, and makes recommendations for further research.
Home smoking restrictions are generally voluntary and may be partial or complete. Partial home smoking restrictions may include smoking only in certain rooms, with a window open, or on special occasions; further, these rules may not apply to guests. In a completely smoke-free home, smoking is not allowed inside by anyone under any circumstances.
The California Tobacco Control Program was one of the first and most successful statewide anti-tobacco campaigns. Much of the program's success may be attributed to changing social norms surrounding smoking behaviours.6–10 While comprehensive population-level interventions are influential in changing societal behaviours, home smoking restrictions have the potential to intervene on a personal level for youth, potentially creating and reinforcing lifelong anti-smoking behavioural values and norms. There is evidence that people who live in jurisdictions with more legislated smoking restrictions (population-level interventions) tend to more frequently have smoke-free homes.5
Parental values and the home environment are influential in the development of youth behaviour, attitudes and beliefs. Parents serve as educators and role models for their children's values and behaviours.9 11 12 Parents may engage in a number of anti-smoking activities to dissuade their children from smoking,12 including having home smoking restrictions, setting limits on free time with peers and establishing anti-smoking contracts.12 In addition, some parents may build their child's resistance skills to prevent them accepting a cigarette offer from peers.12 In our view, establishing a smoke-free home is arguably the most consistent, long-term indication for the child of how tobacco use fits into the parental value system.
This article is a comprehensive review of available studies to date that examine the potential influence of home smoking restrictions on youth smoking behaviours and whether the association appears to vary by parental smoking status, type of restriction (complete or partial) and stage of the smoking uptake continuum. Based on these findings, we make suggestions for future research.
Methods
Article search
We conducted a search of the PubMed database from 1 January 1990 to 26 January 2010 using the following criteria to search all fields: (home OR house OR household) AND (rule OR rules OR ban OR bans OR policy OR policies OR smoke free OR smoke-free OR smokefree) AND (adolescent OR child OR youth OR teen) AND smoking. The only limit on the search was that articles involve those under the age of 18 years. This search yielded 345 articles (figure 1). Of these, we read the abstracts of 49 articles and obtained the full text of 31, of which 17 were found to relate the presence of home smoking restrictions for everyone in the household not just for youth to adolescent smoking behaviour.6 13–28 A further article was identified from the reference lists of the 17,11 and a colleague pointed out an additional article that did not appear in the search results.29 KE, NS and EG read the 19 articles and abstracted the information to appear in table 1. Discrepancies were discussed, and EG made the final decisions regarding the reported classifications.
Derivation of 19 articles reviewed from PubMed search results.
Studies examining the relationship between voluntary smoking policies in the home and youth smoking behaviour
Outcomes
The outcomes examined included: (1) smoking initiation (ever tried a cigarettes or ever smoked 100 cigarettes in lifetime (ever smoker)), (2) status or transitions on the smoking uptake continuum, (3) current smoking (smoke now or in the past 30 days), (4) cigarette consumption among current smokers, (5) intent to smoke, (6) delay in first or daily smoking and (7) smoking cessation among youth ever smokers.
The smoking uptake continuum measures where someone is in the process of moving from being a never smoker to becoming an established smoker. Choi et al31 defined a 12-stage continuum, based on previous smoking behaviour and risk of future smoking. High risk is defined as the lack of a firm intention not to smoke (or smoke again) in the future; it is sometimes called susceptibility to smoking. Adolescents are categorised as: (1) never smokers (low risk or not susceptible to smoking), (2) never smokers susceptible to smoking (high risk), (3) puffers: less than a whole cigarette ever (low risk), (4) puffers (high risk), (5) non-recent experimenters (low risk), (6) non-recent experimenters (high risk), (7) recent experimenters: in past month (low risk), (8) recent experimenters (high risk), (9) former established smoker: at least 100 cigarettes in lifetime (low risk), (10) former established smoker (high risk), (11) current established smoker (low risk) and (12) current established smoker (high risk). In practice, researchers have used fewer categories, particularly at higher stages, and have confirmed such a continuum to be a useful predictor of future smoking.31–35
Results
In 16 of the 19 studies reviewed, there was at least marginal evidence for an association between home smoking restrictions and adolescent smoking behaviour (table 1).11 13 14 16 17 19–21 23–30 Both of the two longitudinal studies, showed a significant relationship.13 20 Because of the importance of this design, these studies are addressed in depth below. Later sections describe various attributes and results from all the studies, with particular emphasis on the role of parental smoking status. Table 1 is organised according to how the studies treat parental smoking.
Longitudinal studies
A longitudinal study (4 years: 2001–2002 to 2005–2006) of 3834 Massachusetts youth (ages 12–17 years at baseline) examined the effect of a smoke-free home on transition from never smoking to experimentation and from never smoking to established smoking (at least 100 cigarettes in lifetime).13 A home was considered smoke free if (1) there were no adult smokers in the home and visitors were not permitted to smoke within the home, or (2) there was an adult smoker in the household but there was a complete ban on smoking inside the home. A three-level hierarchical linear model analysed individual (two levels) and town-level predictors of smoking transitions. Level one individual variables included baseline age, baseline smoking status and having a close friend who smokes; level two predictors were gender, race/ethnicity and household income. Town-level covariates were percentage voting ‘yes’ on a question to increase cigarette taxes and create a statewide tobacco control program, percentage white and percentage younger than 18 years. While overall progression to established smoking was not significantly related to not having a smoke-free home, the tendency to increased transition without a smoke-free home appeared greater among adolescents who lived with a smoker (OR 1.38, 95% CI 0.92 to 2.07) than among adolescents who lived with adult non-smokers (OR 1.08, 95% CI 0.61 to 1.91). The absence of a smoke-free home was associated, however, with transition from never smoking to early experimentation; it was significantly more risky for youth living with non-smokers (OR 1.89, 95% CI 1.30 to 2.70) but was not significant for youth living with smokers (OR 0.88, 95% CI 0.73 to 1.37).
The other longitudinal study interviewed 4255 Minnesota youth ages 11–16 years at baseline, with a 6-year follow-up period.20 The study assessed individual, family and community level variables that were collected every 6 months. The primary outcome variable was smoking in the past month (none vs any) and the primary explanatory variable was presence of a clean indoor air policy at the city or county level that restricted smoking beyond the state-level legislation. A variety of other explanatory factors were present in the model, including a current (at each interview) smoke-free home and parental smoking status at baseline. The study used generalised equation modelling and presented data for unadjusted, minimally adjusted and fully adjusted models. While no overall significant effect was found for city or county clean indoor air policies, a significant association was present in the fully adjusted model for parental smoking that increased adolescents' odds of smoking by 40% (maternal: OR 1.38, 95% CI 1.16 to 1.65, p<0.01), (paternal: OR 1.39, 95% CI 1.16 to 1.68, p<0.01). In addition, a smoke-free home resulted in significantly lower odds of adolescent past month smoking (OR 0.88, 95% CI 0.80 to 0.96, p<0.01).
Differing definitions of home smoking restrictions
While all of the studies assessed some form of home smoking restrictions, they differed in their definition. One study used a multilevel scale based on a number of questions related to smoking in the home,23 and showed a significant association. Seven studies15–17 22 24 29 30 used multiple categories (eg, complete, partial and no restrictions); five of these found a significant relationship for a completely smoke-free home,16 17 24 29 30 but only three found some association for partial restrictions, compared to none.24 29 30
A total of 11 studies11 13 14 18–21 25–28 compared having a completely smoke-free home to all others, and 9 found a significant association.12 17 19 20 25 27 28 33 35 For example, Rissel et al27 found that students in families with clear rules about not smoking in the home were less likely to be current smokers (OR 0.67, 95% CI 0.49 to 0.90). Henriksen and Jackson11 found increased youth experimentation (OR 1.39, 95% CI 1.03 to 1.88) and increased intent to smoke (OR 1.77, 95% CI 1.03 to 1.88) when all homes not smoke free were compared with smoke-free ones.
Smoking rules only one anti-smoking socialisation strategy among others
In seven studies,11 19 21 22 24 27 29 home smoking rules were examined in conjunction with other parental smoking prevention strategies, including communication about smoking, parental warnings about smoking, anti-smoking statements by parents, knowledge about their child's smoking, knowledge about friends smoking, parental psychological control, confidence in influencing child's smoking behaviour and the availability of cigarettes within the home. Six of these studies found evidence that home smoking restrictions were a useful anti-tobacco socialisation strategy after accounting for the others.11 19 21 24 27 29
Type of youth smoking behaviour
Early versus later stage in the uptake continuum
Some studies focused only on the earlier stages of the smoking uptake process,11 19 some only on the later stages,16–18 and some on early and later stages.13 21 23–26 30 All 9 studies that considered the earlier stages found an association between the presence of home smoking restrictions and a more favourable youth smoking status, but only 6 of the 12 studies that considered later stages reported an association.16 17 21 24 26 30 The above summary does not include the five studies that only considered participant smoking behaviour in the previous 30 days,15 20 22 28 29 because it is impossible to separate out regular or later stage smokers from experimenters. Three of these past 30-day studies found an association.20 28 29 One study that considered past 30-day smoking along with other measures also found it to be significant.30
Smoking cessation
Three studies examined cessation.16 17 25 Two of these16 17 used data from the Current Population Surveys and included older youth (15–24 years), and both showed less current smoking with a smoke-free home among those who met the adult definition of an ever smoker (at least 100 cigarettes in lifetime). One of these studies16 reported that the quit ratio (QR) was twice as large for youth in smoke-free homes (QR 28.9 (95% CI 25.2 to 32.6)) when compared to households without any smoking rules (QR 12.5 (95% CI 9.9 to 15.2)). Similarly, the other study17 found more former smokers in smoke-free homes (OR 1.80, 95% CI 1.23 to 2.65) but not for homes with partial restrictions (OR 1.15, 95% CI 0.77 to 1.71) compared to no restrictions. The third study25 included younger adolescents in the sample and did not find a significant association between home smoking rules and not being a current smoker among ever experimenters.
Cigarette consumption
Clark et al16 examined cigarette consumption among youth current smokers and found a significant association between lower consumption and a smoke-free home (OR 0.42, 95% CI 0.24 to 0.71) and partial restrictions (OR 0.97, 95% CI 0.51 to 1.48) versus no restrictions.
Age of initiation
One study addressed the age of first cigarette and age of initiation of daily smoking.26 Using survival analyses, this study found that in homes without smokers or where adult smokers did not smoke inside the home that the risk of early age smoking onset was reduced compared to homes where smokers smoked inside.
Parental/household adult smoking status
The studies differed in how they accounted for parental or other household adult smoking in their analyses. Six studies either included an interaction term between parental smoking and having a smoke-free home23 25 or analysed youth living in households with adult smokers and non-smokers separately.13–15 19 One study that explored an interaction did so as an addition to the main analyses, and the results were only reported in the text.25 Four studies found either an association or a stronger association in families where parents or other household adults did not smoke,13 19 23 25 and one study found a marginal association only in homes with an adult smoker.14
A total of 11 studies included parental or other household adult smoking as a covariate in the multivariate analyses.11 16–18 20–22 24 27 29 30 Of these, nine found home smoking rules were significant when parental smoking was controlled for in the model.11 16 17 20 21 24 27 29 30 In the two negative studies,18 22 the smoke-free home variable was significantly associated with reduced smoking without parental smoking status in the model, but including it rendered the smoke-free home variable non-significant. For instance, in the study by Fisher et al,18 the smoke-free home variable in the model not including adult smoking was significant for reduced youth smoking (OR 0.67, 95% CI 0.48 to 0.93), but when parental smoking was included, this factor was highly significant for increased youth smoking (OR 1.81, 95% CI 1.40 to 2.35), but the variable for a smoke-free home was no longer significant (OR 0.94, 95% CI 0.65 to 1.35). Clearly, these two factors are highly related and their relative prevalence in the sample might have influenced the results.
Two studies were not able to account for parental smoking.26 28 One included only children of smokers and was thus unable to account for parental smoking status.28 It used a path analysis and found a significant indirect effect for a smoke-free home (p<0.0008) through peer smoking. The other study defined the home rule variable in a manner that made it impossible to separate the association of a smoke-free home from the composition of the household regarding adult smoking status.26
Discussion
Taken together, the results of the reviewed studies suggest that home policies may influence youth smoking behaviours, although additional longitudinal studies are required to establish this link. These results also suggest that parental smoking may interact with home smoking restrictions to impact youth smoking behaviour, further reinforcing anti-smoking values of non-smoking parents. In addition, a completely smoke-free home appears to send a stronger anti-smoking message than partial restrictions, and a smoke-free home may be more influential in the earlier rather than later stages of the smoking uptake continuum.
While a smoke-free home policy was more often associated with reduced adolescent smoking in homes with non-smoking parents, the association was less clear in homes with at least one smoking parent or other adult. In some studies that included an interaction term between parental smoking and a smoke-free home, or where smoking and non-smoking homes were analysed separately, a significant association was found mainly in homes where the parents did not smoke. Although the relationship was significant in a few studies for homes where parents smoked, it was considerably weaker. Farkas et al17 reported that when smoking is permitted in homes without adult smokers, youth smoking levels approached the level observed in homes with adult smokers. This suggests that a lack of smoking restrictions, even in homes with parents who never smoked, may weaken communication of parental the anti-smoking values.
As parental smoking status appears related to the effectiveness of home smoking restrictions and adolescent smoking, it is important that future studies include an interaction term between parental smoking and home smoking restrictions or analyse homes with and without adult smokers separately. The former analytic strategy is preferable to preserve statistical power. Failure to appropriately account for parental smoking status in the analysis may result in residual confounding that can lead to misleading results. Also, it is important to account for known explanatory and other factors potentially confounded with smoking restrictions (see table 1).
A smoke-free home appears to be more associated with adolescent smoking behaviour than partial smoking restrictions. The classification of home smoking restrictions that includes total, partial and no restrictions allows total and partial restrictions to be compared to no restrictions. While contrasting a smoke-free home with all others groups (partial and no restrictions together) may reduce the chances of finding a significant association, our review does not substantiate this. Evidence exists that partial restrictions are not as successful in protecting non-smokers, including children, from secondhand smoke.36 37 Thus, a completely smoke-free home should be the goal, and using this categorisation versus all others is probably more relevant.
Imposing a smoke-free home policy when youth are already experimenting with cigarettes may not be as influential as having one in place throughout childhood. In the one longitudinal study that examined different stages in the smoking uptake continuum,13 it was the transition from never smoking to experimentation for which a smoke-free home was significant. Further, a number of the cross-sectional studies that examined earlier and later stages appeared to find more associations of home smoking rules with earlier compared to later stages as the outcome variable. If possible, it might be important to account for the duration of a smoke-free home policy.
Just two studies that examined the impact of home smoking restrictions on youth smoking behaviour were longitudinal.13 20 Cross-sectional designs do not permit researchers to conclude whether the observed associations are causative or simply correlative. While a prospective randomised intervention trial to promote adoption of smoke-free homes would provide more definitive evidence for a smoke-free home to reduce youth smoking, such a study is unlikely for several reasons. To be funded, an intervention study would likely need to include multiple anti-tobacco socialisations strategies, making it difficult to evaluate individually the influence of any one of the tactics. Additionally, as smoke-free homes become more popular, it would be difficult to keep the control group from adopting smoke-free homes. Also, longitudinal studies are not devoid of limitations. Attrition rates are usually considerable. Nevertheless, advanced statistical analytic methods can adjust for variables at different hierarchical levels, time-variant variables and missing data over time.13 20
There is a possibility that there could be an international ‘natural’ experiment in the next few years to address the role of smoke-free homes on reducing youth smoking. As of May 2009, 164 countries were signatories to the Framework Convention for Tobacco Control38 which requires smoke-free workplace legislation. Thus, over the next few years, there is expected to be a marked increase in smoke-free workplace legislation. The history of developed countries over the past decade is that voluntarily imposed smoke-free home policies often increase in the period following increases in smoke-free workplace legislation, although this association may not be causal.5 39 One way that researchers could take advantage of this potential ‘natural’ experiment would be for an exploration of smoke-free homes to be included in the current international surveillance systems.
A limitation of this review is that most studies were undertaken in developed countries, particularly the USA, and there is less certainty about the generalisability of these findings to other cultures with different economic systems, social and familial structures and ethnic identities and traditions . As with any review of this nature, there is the possibility that negative studies were not accepted for publication. All studies determined smoke-free home status by respondent report. Ideally, future studies could include a smoke monitor in the home or a biomarker (eg, cotinine) to provide a more objective measure that the home was indeed smoke free. In our review, we did not include studies where smoking restrictions applied to youth but may or may not have applied to adults. Rules against smoking in the home that apply only to the youth may be antithetical to the promotion of anti-tobacco values. Such a policy implicitly sends the message that it is not appropriate for youth to smoke, but that it is an acceptable adult behaviour. Youth wanting to emulate adult behaviour may be more likely to smoke. This strategy of presenting smoking as a valid adult behaviour was employed by the tobacco industry in a youth smoking prevention media campaign, which appeared to result in increased adolescent smoking.40
In summary, we recommend that studies planned to evaluate future ‘natural experiments’ be longitudinal with multiple survey interviews over time, employ hierarchical analytic strategies, contrast smoke-free homes against all others, properly account for parental or other household adult smokers and other important explanatory/confounding variables, and look at transitions early and later in the smoking uptake continuum.
References
Footnotes
Funding This work was supported by the Tobacco-Related Disease Research Program grant 15RT-0238 from the University of California. Some of the work for this review was undertaken by one of the authors (EAG) as part of the preparation of a chapter for a handbook on smoke-free policies for the IARC.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.