Trends in Second-Hand Tobacco Smoke Exposure Levels at Home among Viet Nam School Children Aged 13-15 and Associated Factors

Second-hand tobacco smoke (SHS) exposure at home, especially among children, is a serious issue in Viet Nam. During the past decade, much effort has been taken for tobacco control in the country, including various prgorammes aiming to reduce SHS exposure among adults and children. This article analysed trends and factors associated with SHS exposure at home among school children aged 13-15 in Viet Nam, using the Global Youth Tobacco Surveys conducted in 2007 and 2014. Descriptive and inferential statistical methods with logistic regression were applied. Overall, there was a significant reduction in the level of exposure, from 58.5% (95%CI: 57.6-59.3) in 2007 to 47.1% (95%CI: 45.4-48.8) in 2014. Of the associated factors, having one or both parents smoking was significantly associated with the highest odds of SHS exposure at home (OR=5.0; 95%CI: 4.2-6.1). Conversely, having a mother with a college or higher education level was found to be a protective factor (OR=0.5; 95%CI: 0.3-0.8).


Introduction
Second-hand tobacco smoke (SHS) exposure can cause a long list of diseases; reducing SHS exposure is one of the 6 key measures to control tobacco-related harm, as recommended by the World Health Organization (WHO, 2008).
SHS exposure can happen in both public places and at home. A recent published factsheet (CDC, 2016) based on the Global Youth Tobacco Survey (GYTS) 2007 showed a very high level (58.5%) of SHS exposure at home in Viet Nam. Since then, significant progress has been made in Viet Nam in tobacco control, including the approval of the National Tobacco Control law in 2012, which is in line with the WHO FCTC.
This article examines levels of SHS exposure at home among school children aged 13-15 between 2007 and 2014, with Asia-Pacific countries, and socio-demographic factors associated with the level of SHS exposure at home.

Materials and Methods
The GYTS survey was conducted using a standardized 1 questionnaire, with data cleaning and weighting support from CDC. The description of these surveys can be found (Giang et al., 2016) in the same special issue of this journal. The age eligibility for inclusion in this study is 13-15 years old. The dependent variable is the level of SHS exposure at home, which is defined as "any student who witnessed a person smoking inside the home in his/her presence, in the past 7 days before the survey." Independent variables included: current smoker, defined as respondents who smoked at least one cigarette, or another tobacco-smoking product, during the last 30 days; parental smoking, defined as having a parent or both currently smoking; weekly money, the allowance given by parent, was coded into 4 groups-no weekly money, less than 20000 VND, 20000 to 50000, and more than 50000. For the 2007 survey, the unit was monthly money, which we have transformed into weekly money; father and mother education is recorded into 4 groups: less than primary, primary and secondary school, high school, and college and higher; father and mother occupation is recoded into three groups: farmers (including farmers, fishing and forestry), government employees, and others; region of Viet Nam is classified under the three major regions, the north, central, and south. Some variables available in the 2014 survey, but missing from the 2007 survey include region, parent education level and parent occupation. Therefore, no comparisons could be made for these categories between the two surveys.
The data analysis was conducted using Stata 10 statistical software. Chi-square was used to test for differences in the prevalence of SHS exposure between 2007 and 2014. Statistical analyses were conducted using both descriptive and inferential methods. Univariate and Multiple logistic regressions were used to explore the effects of different socio-demographic factors that impacted on level risk of SHS exposure at home.
Ethical considerations: The study was presented for review by the Ministry of Health. In addition provincial authorities granted legal permission to proceed with the research. Furthermore, prior to conducting the survey, teachers, parental representatives, and students provided consent.

Results
The GYTS survey 2007 had a total of 15,610 respondents from 134 schools located in 9 provinces of Viet Nam. After excluding those outside the age range of 13-15, the number eligible for analysis was 14,637 students. Of these, boys and girls accounted for 45.5% and 54.5%, respectively.
The GYTS 2014 had a smaller respondent of 3553 respondents from 50 schools, located in 13 provinces of Viet Nam. The number of students eligible for analysis, aged 13 to 15, was 3430, of which 49.1% were boys and 50.9% were girls.
The results on trend of SHS exposure are summarized in table 1. There was a significant reduction in the level of overall SHS exposure at home from 58.5% in 2007 to 47.1% in 2014 (p<0.01). The reduction was evenly distributed in all socioeconomic groups between the two iterations of the surveys, all at a level of high statistical significance (p<0.01). Socio-economic factors associated with higher odd of SHS exposure are presented in table 2. There were significantly higher odds of SHS exposure for some subgroups, while for other subgroups there were no significant differences. There were some differences in univariate and multiple logistic regression results. For example, the current smoker group is not significantly different in univariate regression, OR=1.6 (95%CI: 0.98-2.6), but does yield a statistically significant variability in the multiple regression model, OR=1.7 (95%CI: 1.1-2.8). On the other hand, some variables, such as father having college or university education is significantly different from the reference group in univariate analysis with OR=0.5 (95%CI: 0.4-0.7), but becomes statistically insignificant in multiple regression model, OR=1.0 (95%CI: 0.6-1.6). Similar patterns are noticed for groups with father or mother holding governmental positions. Lastly, the group with a mother having a college or higher education had significantly higher odd in both univariate and multiple regression analyses, with OR=0.5 (95%CI: 0.4-0.6) and OR=0.5 (95%CI: 0.3-0.8) respectively.
For the final results, we will use the results of the multiple logistic regression models for its superiority over the univariate model. The group having the highest   risk were those having a parent smoking currently, with 5 time greater odd of exposure (OR=5.0; 95%CI: 4.2-6.1). A similar trend, though to a lesser degree, were found those respondents who were current smokers, which was associated with a 1.7 higher odd of exposure (OR=1.7; 95%CI: 1.1-2.8).
Conversely, respondents with a mother having a college education or higher had a significantly lower odd of SHS exposure at home; it was half compared to other groups (OR=0.5; 95%CI: 0.3-0.8). It is interesting to note that a father's higher education did not have any protective impact on reducing odd of exposure to their children (OR=1; 95%CI: 0.6-1.6).

Discussion
From the results of the 2014 GYTS survey, there is evidence of significant reductions in the SHS exposure levels at home among school children, from 58.5% in 2007 to 47.1% in 2014. This reduction is expected given significant investment, effort, and progress made in tobacco control in Viet Nam in the past few years, especially the adoption of a strong tobacco control law in line with the FCTC and several mass media communication campaigns with funding and technical support from the World Lung Foundation, WHO, Bloomberg Initiative partners, and others (WLF, 2015).
Besides the GYTS 2007, there was a National Global School Health Survey conducted in Viet Nam in 2013 that also asked about SHS exposure. The survey, however, did not distinguish between exposures at home or at public places. In addition, the age groups were also different. There was no other national survey covering SHS exposure among this group of school aged children.
As seen in table 3, when comparing SHS exposure levels among selected countries in the Asian region using GYTS data from the CDC Global Tobacco Surveillance System Data (CDC, 2016), Vietnam's exposure levels in the 2007 survey (58.5%) was the second highest. The level of exposure in the 2014 survey (47.1%) was significantly reduced when compared to 2007, but still significantly higher than most other countries, with the exception of 2009 data from Malaysia and Indonesia (which is data that is now 8 years out-dated).
Among the sub-groups, those who have one or both parents currently smoking suffered the highest odd of SHS exposure at home (OR=5.0). This is logical and also in agreement with evidence from GYTS studies elsewhere (Peltzer, 2011;Raute, 2012). The level of a mother's education, especially a mother having college or higher education clearly had a positive impact in reducing SHS exposure levels. However, a father's education level seemed to have no effect. This may be explained by two suppositions: 1) the rate of adult women who smoke (according to the GATS 2010) is only 1.5%, suggesting most mothers aren't smokers and 2) culturally, mothers in Viet Nam have a higher tendency to be proactive about protecting their children's health. Conversely, males are more likely to smoke (almost one in two adult males are smokers according to GATS 2010), so they may be less proactive about taking specific steps to reduce SHS exposure to their children.
The positive association between higher levels of SHS exposure and having pocket money to spare, a proxy for higher income families, seemed to be counter-intuitive. When cross-referenced with the 2007 survey, there was a similar trend: those who have no pocket money have less SHS exposure. When compared with other countries, this finding is unusual and surprising. A review by (Orton, 2014) found that most studies (10 out of 11) showed significantly higher risk of SHS exposure among those with lower socio-economic classes. One potential explanation is that children with no pocket money tend to reside in rural areas, where there is an open court-yard in nearly all households and children spend more time outdoors. These varying conditions may explain the lower level of SHS exposure inside the home. It is unfortunate that we do not have a variable for rural versus urban groups, and therefore could not test this hypothesis.
It is noted that this study is not without limitations. Although, both surveys in 2007 and 2014 were considered nationally representative surveys, there were notable differences in the sampling methods and sample size of the two surveys. Therefore, comparison of the results with those of the earlier study should be interpreted with suitable level of caution.
In conclusion, there has been significant progress in reducing SHS exposure to school children aged 13-15 in Viet Nam, when comparing results from 2007 to 2014, suggesting that tobacco control efforts are in the right direction. Still, the level of exposure, even after a sizable reduction, is still much higher when compared with other countries in the region. Therefore, greater effort is needed to increase awareness among smokers, typically adult males, to encourage smoking outside of homes as well as smoking cessation efforts in order to reduce the effects of SHS exposure.
It is recommended that future GYTS studies also collect information for the variable of urban vs. rural locality and the sampling method should be kept stable in the future rounds.