Article Text
Abstract
Objective To investigate the effect of comprehensive smoke-free legislation in 2007 on the exposure of children to secondhand smoke (SHS) in Hong Kong.
Methods Two cross-sectional questionnaire surveys were conducted, before (2006) and after (2008) the implementation of smoke-free legislation, among primary 2–4 students (equivalent to US grades 2–4) from 19 and 24 randomly selected schools, respectively. Adjusted ORs for SHS exposure at home and outside home post-legislation compared with pre-legislation were calculated. The strength of the association between SHS exposure and respiratory symptoms in each survey was used as an indirect indicator of the intensity of exposure.
Results Among 3243 and 4965 never smoking students in the 2006 and 2008 surveys, the prevalence of SHS exposure in the past 7 days increased both at home (from 10.2% to 14.1%) and outside home (from 19.8% to 27.2%). Post-legislation, students were 56% more likely (p<0.01) to report SHS exposure at home coupled with an insignificantly stronger association between SHS exposure and respiratory symptoms. Similarly, students were 60% more likely (p<0.001) to report SHS exposure outside home in 2008, but the association between SHS exposure and respiratory symptoms became insignificantly weaker. Parental smoking rates were similar before and after legislation.
Conclusions The prevalence of exposure to SHS at home and outside home have both increased among primary school students in Hong Kong post-legislation. Comprehensive smoke-free legislation without strong support for smoking cessation might have displaced smoking into the homes of young children.
- Environmental tobacco smoke
- prevalence
Statistics from Altmetric.com
Introduction
Worldwide, more than 43% and 55% of youths are exposed to secondhand smoke (SHS) at home and in public places, respectively.1 SHS causes many diseases2 including respiratory symptoms in children.3–5 We have shown that children who were exposed to SHS were 54% more likely to report respiratory symptoms compared with unexposed children.6 Despite the tremendous health burden caused by SHS, there is little evidence that family-based or school-based interventions are effective in reducing SHS exposure in children.7
Smoking bans in public places were found to reduce SHS exposure in non-smoking adults8 and bar workers.9 However, there are concerns that public smoking bans may displace smoking into homes, increasing children's exposure to SHS. Only a few studies, all of which are British, have described the effects of smoke-free legislation on smoking in the home and SHS exposure in children.10–12 In Ireland, smoke-free legislation did not increase cigarette consumption at home.10 In England, in line with the drop in parental smoking rates, the cotinine concentration in children has almost halved from 1988 to 1998, but the reductions were limited to non-smoking households, and hence exposure outside home.11 Similarly, saliva cotinine in children dropped by 39% post-legislation in Scotland overall, but the effect was mainly seen in non-smoking and father-smoking-only families.12 However, a recent study found a significantly higher serum cotinine level among non-smoking adults in New York City than the national average. Given that New York City had a lower smoking prevalence and probably the strongest smoking ban in public places in the USA, this finding suggests that smoke-free legislation in a densely populated city might increase SHS exposure.13 Therefore, the displacement of smoking from public places to the family home remains an important public health issue that warrants further investigation.
In Hong Kong, substantial amendments to the smoke-free legislation became effective in January 2007 (table 1). Previously, smoking had been banned in public transport carriers, shopping malls, hospitals and partially banned in restaurants with more than 200 seats. The new ordinance further covered all indoor eating places, workplaces, shops, markets, playgrounds, escalators, beaches and most parks. However, some pubs and adult entertainment premises were exempted from the smoking ban until July 2009. In addition, the housing department banned smoking in all common (including outdoor) areas of public housing estates where rents are heavily subsidised as a public welfare and almost half (46.8%) of the population live.14 Penalty points are allotted to households for smoking and other offences, with the ultimate punishment being the termination of tenancy.
The new tobacco control measures were widely publicised and generally welcomed by the public.15 We have previously reported that 85% of adults in Hong Kong were exposed to SHS in restaurants in the previous 30 days in 2000.16 The corresponding prevalence in children was conservatively estimated at 35% (85 × 41%) as 41% of adults in the same study reported eating out with their children during the past 7 days.16 Under the new ordinance, children's exposure to SHS in eating places was expected to be greatly reduced but there were concerns that smokers who did not quit would resort to smoking at home or on the streets, especially those who live in public housing estates.
The new legislation is quite effectively enforced by multiple government departments coordinated by the Tobacco Control Office. In 2007 and 2008, there were 3780 and 7305 summonses against smoking offences, respectively.17 Hong Kong is densely populated (6349 people per km2), typically with narrow pavements. Likewise, children generally live in small flats sharing a bedroom with others. Any change in smoking in the family home and public places as a result of the smoke-free ordinance would probably affect the exposure of children to SHS. We therefore evaluated the effects of the legislation on the exposure of children to SHS at home and outside home in Hong Kong.
Methods
Two cross-sectional school-based surveys were conducted among primary 2–4 (P2-4, equivalent to US grades 2–4) students before (January to March 2006) and after (January to March 2008) the implementation of smoke-free legislation in January 2007. The schools were selected from those which had enrolled to receive an educational theatre stage performance organised by the Hong Kong Council on Smoking and Health. Each year, all primary schools were invited to enrol and the first 100 were accepted. In the 2006 survey, 19 primary schools were randomly selected from the 100 enrolled schools. In 2008, 14 of the 19 schools were again included, and another 10 schools were randomly selected from the remaining enrolled schools, totalling 24 schools. The findings based on the 14 or 24 schools were similar (table not shown) so that all of the 24 schools in 2008 were used in the analyses.
The questionnaires were similar in the two surveys and they were anonymous, self-administered and written in simple Chinese. SHS exposure at home and outside home was measured by two questions separately: ‘How many days in the past 7 days were you exposed to secondhand smoke at home?’ and ‘How many days in the past 7 days were you exposed to secondhand smoke outside home?’ The eight response options from 0 to 7 days/week were recoded as ‘none’ ‘1–3 days/week’ and ‘4–7 days/week’.
Self-reported exposure to SHS in 2006 by 16 randomly selected non-smoking students was validated against hair nicotine concentration by the Johns Hopkins University. SHS exposure was considered positive for nicotine concentration exceeding 0.07 ng/mg.18 The validation results were satisfactory with sensitivity of 75%, specificity of 63%, positive predictive value of 67% and negative predictive value of 71%.
In 2008, the students additionally reported the location of SHS exposure outside home in the past 30 days with options including streets, parks, restaurants and others. A longer time frame of 30 days was used to obtain some estimates for places where SHS exposure were likely to be less common such as shopping malls and playgrounds. Students who reported frequent cough or phlegm were classified as having respiratory symptoms. Information on smoking behaviour, parental smoking status and demographic characteristics was also obtained.
SPSS 16.0 and Stata 9.2 were used for data analysis. Descriptive data were weighted by census school grade distribution in 2006 and 2008. The prevalence of SHS exposure was compared between the 2006 and 2008 surveys, and adjusted odd ratios (ORs) for exposure in 2008 compared with 2006 were calculated using logistic regression adjusting for sex, age, school grade, place of birth, parental smoking status and the clustering effect of schools.
Apart from the frequency of exposure, the intensity of exposure to SHS also depends on other factors such as the duration and concentration of SHS exposed. Such information is, however, difficult to report, especially by young children. Since respiratory symptoms are closely linked to SHS exposure in children, they were used as an indirect indicator of the intensity of SHS exposure. To eliminate the effect of active smoking on respiratory symptoms, ever smokers in 2006 (4.6%) and 2008 (4.7%) were excluded from the analysis.
In each survey, logistic regression was used to calculate respiratory symptom ORs for SHS exposure at home and outside home adjusting for each other, sex, age, school grade, place of birth, parental smoking status and school effect. Mean air pollution index and the population influenza consultation rate over 3 months and 1 week, respectively, preceding the survey date of each student were additionally adjusted in sensitivity analyses. The interaction effects between the year of survey and SHS exposure in relation to respiratory symptoms were estimated using interaction terms of year multiplied by SHS exposure.
Invitation letters were sent to parents and students for passive consent. Only those parents who declined participation needed to return the signed reply forms. Ethical approval was obtained from the institutional review board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster.
Results
The response rates of the 2006 and 2008 surveys were 68% and 83% at the school level, and 96% and 93% at the individual level, respectively. A total of 3243 and 4965 never-smoking P2-4 students in 2006 and 2008, respectively, were included in the analyses. The 2006 survey has 50.3% boys and a mean age (SD) of 8.3 (1.0) years; the 2008 survey had 52.2% boys and a mean age (SD) of 8.6 (1.2) years. In Hong Kong, the place of birth (Hong Kong, Mainland China and other places) is associated with socioeconomic status in children.19 We found a similar distribution of the place of birth between the 2006 and 2008 samples (p for χ2=0.88). The five schools that dropped out were also similar to the follow-up schools in terms of sex (p for χ2=0.65), age (p for t test=0.90), place of birth (p for χ2=0.35) and parental smoking status (p for χ2=0.36).
Table 2 shows that the prevalence of exposure to SHS increased from 2006 to 2008 (all schools) both at home (from 10.2% to 14.1%) and outside home (from 19.8% to 27.2%). Regardless of the place of exposure, 23.2% and 31.2% of students were exposed to SHS in 2006 and 2008. The prevalence of any parental smoking was similar in 2006 (36.1%) and 2008 (35.0%). The prevalence of respiratory symptoms increased slightly from 36.6% in 2006 to 38.5% in 2008. The most common places of SHS exposure outside home in 2008 were the streets (34.4%), parks (15.4%), restaurants (13.4%) and bus stops (12.7%). The results based on all schools and follow-up schools in 2008 were similar.
Table 3 shows that after adjusting for potential confounders, students in 2008, compared with 2006, were 51%, 58% and 56% more likely to report SHS exposure at home for 1–3 days, 4–7 days and any days, respectively (all p<0.01). The corresponding figures for SHS exposure outside home were 45%, 101% and 60%, respectively (all p<0.01). Overall, students were 54% more likely to report any SHS exposure at home or outside home in 2008 than 2006 (p<0.01).
Table 4 shows that exposure to SHS at home for 4–7 days per week was significantly associated with respiratory symptoms with an adjusted OR (95% CI) of 1.19 (1.00 to 1.50) in 2008, compared with 1.09 (0.79 to 1.50) in 2006. The each-day increase in home exposure was associated with 7% excess risk of respiratory symptoms in 2008 but no increased risk in 2006. On the other hand, the association between exposure to SHS outside home for 4–7 days and respiratory symptoms seemed to be weaker in 2008 with OR of 1.54 (1.20 to 1.98) compared with that of 2.06 (1.34 to 3.18) in 2006, although the difference was insignificant. Moreover, the respiratory symptom risk for the each-day increase in SHS exposure outside home decreased from 35% in 2006 to 20% in 2008. Regardless of exposure at home or outside home, the associations between SHS exposure and respiratory symptoms were quite similar in the two surveys. The results were similar after additionally adjusting for air pollution and the influenza consultation rate (data not shown).
Discussion
This is the first report on the effects of smoke-free legislation in children in a non-Western setting. One year after the implementation of legislation, we found a significant increase in the prevalence of SHS exposure in children both at home and outside home (table 2) using a repeated cross-sectional study design. After adjusting for potential confounders, the odds of exposure post-legislation at home and outside home were significantly higher by 56% and 60%, respectively. Larger increases in the prevalence of exposure were observed for frequent exposures of 4–7 days per week than 1–3 days per week (table 3). Although residual confounding due to unmeasured background differences in the two survey samples, such as parental smoking frequency and consumption, could not be ruled out completely, any such effect is unlikely to be large as the schools in both surveys were similar.
As parental smoking rates were stable in the two surveys (table 2), the increased prevalence of exposure at home was probably due to increases in the frequency and/or duration of smoking in the home by parents and other family members. This increase in the prevalence of exposure was apparently accompanied by a strengthening of the association between home exposure to SHS and respiratory symptoms post-legislation, although the difference was statistically insignificant (table 4). These findings confirmed local anecdotal reports that smoking was displaced into homes after being banned in most public places.20 Given the typical urban high-rise living in Hong Kong compared with houses and gardens in Western countries, it would be difficult for young children to avoid being exposed to SHS if someone smokes at home. Although we did not collect data on housing type, such smoking displacement had probably been more common in public housing estates because of their stronger smoking bans in common areas.
A similar rise in the prevalence of exposure to SHS outside home was reported post-legislation (table 2). After adjusting for potential confounders, children in 2008 were 60% more likely to report any SHS exposure outside home compared with 2006 (table 3), which seemed to conflict with the legislative measures that banned smoking in public places. Based on our previous study, it was estimated that 35% of children were exposed to SHS in restaurants in the past 30 days.16 One year post-legislation, this has reduced to only 13.4% (table 2), which indicated a major success of the legislation and at the same time the need for more stringent enforcement of the smoking ban. The reduction of SHS exposure in restaurants to a low level was consistent with the sporadic smoking by individual offenders observed in restaurants, who usually would comply after being dissuaded by waiters without police intervention. The improvement in air quality in restaurants post-legislation was generally well recognised and welcomed by the public.
Streets were by far the most common (34.4% in the past 30 days) location of SHS exposure outside home in 2008, followed by parks (15.4%). However, both of these exposures were outdoors and likely to be momentary in nature. Apart from bus stops (12.7%), other locations of exposure to SHS outside home were reported by less than 10% of students (table 2). Therefore, although the overall prevalence of exposure to SHS outside home had increased, it was mainly due to short exposures outdoors while the previous major source of exposure in restaurants had reduced. This was supported by the apparently weaker association between SHS exposure outside home and respiratory symptoms post-legislation (table 4).
Regardless of the locations at home or outside home, P2-4 students in Hong Kong were 54% more likely to report exposure to SHS post-legislation (table 3), although the strength of association between such exposure and respiratory symptoms was similar (adjusted OR of 1.18 in 2006 and 1.14 in 2008 for each day increase in exposure). This suggested that among children who were exposed to SHS, the intensity of exposure per day was similar before and after the implementation of smoke-free legislation.
Although the displacement of smokers into the homes has always been a concern for the implementation of smoke-free legislations in different countries, such evidence has never been reported to the best of our knowledge. Our findings unfortunately showed that this has happened in Hong Kong among some families with primary school children. The crowded housing and outdoor environment were partly to blame but the lack of complementary support for population-wide smoking cessation such as that in the UK was probably another, if not more important, contributing factor.
The legislation, which also banned smoking in all indoor workplaces, has probably benefited working adults more than young children. Parents who smoke should be specifically targeted to quit smoking or at least to avoid exposing their children to SHS. The tobacco tax was recently increased by 50% in Hong Kong; part of this revenue should be used to support smoking cessation services for all smokers. SHS is a class I human carcinogen, exposing young children to SHS, particularly at home, is tantamount to physical abuse and should be forbidden. The legal and ethical issues involved are complex and controversial,21 as they were a few decades ago when banning smoking in public places was first discussed. Having established the harm of SHS, we should act immediately and decisively to protect children from further exposure to SHS.
This study has several limitations. Exposure to SHS was based on the self-reported number of days exposed in the past 7 days rather than biomarkers such as cotinine, although self-report is a standard approach as used in the Global Youth Tobacco Survey.22 Children's self-reported SHS exposure was highly correlated with that measured using biomarkers.23 Our validation study using hair nicotine in 2006 also found satisfactory test results for the self-reported data. Using self-reported SHS exposure also allowed us to estimate the prevalence of exposure at home and outside home separately.
Without objective measures of SHS exposure, we used the strength of the association between SHS exposure and respiratory symptoms as an indicator of the intensity of exposure. Individuals with vastly different age and health conditions may have different opinions of what constitutes SHS exposure and respiratory symptoms, but the similarity of subjects in the two surveys should have largely avoided this problem. Moreover, the two surveys were conducted during the same months in 2006 and 2008 to avoid seasonal differences, and we have considered the effects of air pollution and influenza on respiratory symptoms. Schoolchildren are capable of reporting their health conditions reliably24 and the presence of frequent cough and phlegm should be quite obvious to the students. The greater ORs for respiratory symptoms associated with more frequent SHS exposure and the comparability of ORs with those observed in our previous studies in children2 25 also support the validity of this approach. As we did not measure the place of SHS exposure in the 2006 survey, the apparently low prevalence of SHS exposure in restaurants and high prevalence in streets post-legislation needs to be interpreted cautiously. Finally, the possible increased awareness of the public towards SHS after the legislation might have led to over-reporting of SHS exposure. However, further adjustment of the students' knowledge and attitude towards SHS exposure had little effect on our findings (tables not shown).
Conclusions
The prevalence of exposure to SHS at home and outside the home have both increased among primary school students in Hong Kong post-legislation. However, the intensity of SHS exposure outside the home was apparently attenuated in 2008 with a large reduction of exposure in restaurants, and outdoor exposures being the most common. Our findings indicated that the comprehensive smoke-free legislation of Hong Kong in 2007 might have displaced smoking into the homes of young children. Comprehensive smoking bans in public spaces need to be accompanied by strong support for smoking cessation to avoid displacement of smoking into the homes, especially for populations living in crowded urban environments typical of large cities in South-East Asia and China. Ultimately, smoking near young children, particularly at home, should be banned.
What this paper adds
Smoking bans in public places have raised concerns that they may displace smoking into the homes of young children. Such displacement of smoking was not evident in Western populations but little was known about this in Asian populations.
In densely populated Hong Kong a comprehensive smoking ban in indoor public places without strong support for smoking cessation might have displaced smoking into homes, resulting in higher prevalence of exposure to secondhand smoke among young children.
Acknowledgments
We thank the Hong Kong Council on Smoking and Health for funding support.
References
Footnotes
Competing interests None.
Ethics approval This study was conducted with the approval of the institutional review board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster.
Provenance and peer review Not commissioned; externally peer reviewed.