Article Text
PDF
Abstract
Objective To provide the first analysis of socioeconomic inequalities in children's daily exposure to indoor smoking in households in 26 low-income and middle-income countries (LMICs).
Methods We used nationally representative household samples (n=369 654) collected through the Demographic Health Surveys between 2010 and 2014 to calculate daily exposure to secondhand smoke (ESHS) among children aged 0–5 years. The relative and absolute concentration (RC and AC) indices were used to quantify wealth-based inequalities in daily ESHS in each country and in urban and rural areas in each country. We decomposed total socioeconomic inequalities in ESHS into within-group and between-group (rural–urban) inequalities to identify the sources of wealth-based inequality in ESHS in LMICs.
Findings We observed substantial variation across countries in the prevalence of daily ESHS among children. Children's ESHS was higher in rural areas compared to urban areas in the majority of the countries. The RC and AC demonstrated that daily ESHS was concentrated among poorer children in almost all countries (RC, median=−0.179, IQR=0.186 and AC, median=−0.040, IQR=0.055). The concentration of ESHS among poorer children was greater in urban relative to rural areas. The decomposition of the overall socioeconomic inequality in daily ESHS revealed that wealth-based differences in ESHS within urban and rural areas were the main contributor to socioeconomic inequalities in most countries (median=46%, IQR=32%).
Conclusions Special attention should be given to reduce ESHS among children from rural and socioeconomically disadvantaged households as social inequalities in ESHS might contribute to social inequalities in health over the life course.
- Low/Middle income country
- Secondhand smoke
- Socioeconomic status
- Disparities
Statistics from Altmetric.com
Introduction
Tobacco use is a major global health challenge that accounts for 6 million deaths each year (10% of all adult deaths) and ranks among the most preventable causes of death in the world.1 ,2 Among many of the health risks associated with tobacco use is the harm caused by exposure to tobacco smoke and attendant respiratory irritants, also called secondhand smoke or passive smoking, which affects the health of smokers and those around them, especially children.3 The detrimental health effects of exposure to environmental tobacco smoke are a particularly substantial public health challenge in low-income and middle-income countries (LMICs), where ∼80% of smokers live worldwide.1 In contrast to higher income countries, the absolute number of daily smokers has increased remarkably in LMICs over the past 35 years due to population growth4 and increased marketing activities of tobacco companies in these countries,5 where tobacco control regulations are not yet comprehensive.6
Exposure to secondhand smoke (ESHS) increases risk for several serious cardiovascular and respiratory diseases, including lung cancer and coronary heart disease, in adults.1 ,7 Children may be particularly vulnerable to the adverse effects of ESHS because they have lesser control over their indoor environments and experience greater ESHS compared to adults.8 Moreover, since they are still developing physiologically, children may be particularly sensitive to ESHS;3 secondhand smoke exposure among children can cause asthma and increases the risks for sudden infant death syndrome, respiratory tract infections (eg, pneumonia and bronchitis) and middle ear infections.8 Exposure to passive smoke may also have indirect effects on children's health because household expenditures on tobacco products compete with spending on food, healthcare and other basic resources;9 these indirect effects may particularly be pronounced for children living in low socioeconomic status (SES) households.
In spite of the health and economic consequences of smoking and secondhand smoke exposure, nearly half of children breathe smoke-polluted air on a regular basis and >40% of children in the world live in a home with at least one parent who smokes.1 There has been extensive research about smoking and its harmful effects, including the impact of ESHS on the health of non-smokers,10–12 leading to policies that prohibit smoking in designated workplaces and public venues. However, these regulations do not directly affect smoking behaviours within the home, where children's ESHS is most likely to occur.13
There is a limited literature that focuses on levels of ESHS in households and the range of exposures among women and children.14 It is likely that there are social gradients in children's ESHS because smoking rates are generally higher in socioeconomically disadvantaged households.15 Prior work suggests that rates are especially high among the long-term unemployed, homeless, mentally ill, prisoners, single parents and some groups of new immigrants and ethnic minorities.16 This implies that children living in socioeconomically disadvantaged households are at higher risk of ESHS because their parents or other household members are more likely to smoke.17 In addition, there are fewer smoking restrictions in their homes.18
To date, extant studies from LMICs19 ,20 and high-income countries21–24 have consistently documented an inverse relation between SES and smoking status. Moreover, some recent work25–30 has analysed socioeconomic gradients in ESHS among children in higher income contexts. The relation between SES and ESHS among children in LMICs, however, remains poorly understood. This study aims to provide the first comprehensive analysis of socioeconomic inequality in daily ESHS among children in LMICs. We restricted our analysis to children aged 0–5 years because infants and young children spend more time at home and therefore have more ESHS. In addition, ESHS may have a more adverse effect among children due to their early developmental state.31–33 We used the Demographic Health Surveys (DHS) to report wealth-based inequalities in children's exposure to indoor smoking in 26 LMICs using recently collected information on daily household ESHS. Moreover, since we observed significant rural–urban differences in the prevalence of daily ESHS among children in the majority of countries, we decomposed the overall wealth-based inequality in daily ESHS in each country into inequality within and between urban and rural areas. Measuring socioeconomic inequalities in ESHS among children and the determinants of such inequalities can help guide strategies to reduce the harmful effects of passive smoke exposure for children in LMICs.
Methods
Data
Our analyses were based on data collected through the MEASURE DHS project. The project is funded by the US Agency for International Development (USAID) and receives contributions from other donors and some financial support from participating countries to conduct surveys. The data for our study were obtained from the Standard DHS from 26 LMICs. The Standard DHS are cross-sectional and nationally representative household surveys with large sample sizes (usually between 5000 and 30 000 households), typically conducted about every 5 years in selected LMICs.34 ,35
The DHS uses a multistage sampling procedure36 to obtain comparable information on a wide range of health topics,37 including ESHS in the household. Key advantages of the DHS are high response rates and national coverage.34 Face-to-face interviews by well-trained interviewers, standardised tools and methods, and a similar core set of survey questions are used to facilitate standardisation and comparability of surveys across time and countries.38 ,39 Detailed information on data collection methods, reliability assessment and validation can be found elsewhere.37 We used information collected through the Household Questionnaire from 26 countries surveyed over the period between 2010 and 2014 because the DHS started to collect information on ESHS in 2010. The Household Questionnaire collects general information of all household members (eg, age, sex, relationship to the head of the household, education and birth registration) and household characteristics such as the source of drinking water, toilet facilities, cooking fuel, assets of the household and ESHS.40 The interviewer identifies a capable adult member of the household to serve as the main respondent. Any adult member of the household member can fill out the questionnaire and other members may be consulted to obtain particular information.41 The overall sample comprised of 1 810 017 household members from 369 654 households. After we restricted our observations to children aged 0–5 years, our final sample for analysis included 313 857 children.
Measures
The outcome variable of interest in the study, daily ESHS, was measured using a question in the DHS that ask about the frequency that household members smoke inside the house (ie, never, daily, weekly, less than monthly and monthly). We calculated socioeconomic inequalities in daily ESHS among children aged 0–5 years using a calculated wealth index (WI) for each DHS. Using easy-to-collect data information on a household's ownership of selected assets (eg, televisions and bicycles), types of sanitation facilities and water source and materials used for housing construction, the DHS uses a method suggested by Filmer and Pritchett42 to construct the WI.43 We examined country-level patterns of socioeconomic inequalities in daily ESHS according to socioeconomic indicators collected from the World Bank's World Development Indicators and Global Development Finance.44 The gross domestic product (GDP) per capita (purchasing power parity and current international dollar) and adult (+15) literacy rate were used as indicators of country-level development.
Statistical analysis
To examine the socioeconomic gradient of ESHS among children, we first used the relative and absolute concentration (RC and AC) indices to measure the degree of socioeconomic inequality in ESHS prevalence in our selected countries. We then decomposed total socioeconomic inequalities in ESHS into within-group and between-group (rural–urban) inequalities to identify the sources of socioeconomic inequality in ESHS.
Measuring socioeconomic inequalities
The concentration index approach was employed to measure wealth-related inequality in ESHS among children in each country. The RC index is based on the (standard) concentration curve and can be used to quantify the degree of socioeconomic inequality in a health outcome. The (standard) concentration curve is constructed by plotting the cumulative percentage of the sample population, ranked in ascending order by a measure of SES such as household wealth, against the cumulative percentage of the health-related outcome variable of interest (ie, ESHS). The RC index is determined as twice the area between the (standard) concentration curve and a 45° line representing perfect equality. The RC can be computed using a convenient formula that defines the RC in terms of the covariance between the fractional rank of individuals in the SES distribution and their health-related variable as follows:45–47
1where Yi is individual (child) i's health-related variable of interest (ie, ESHS), µ is the mean of the health variable for the total sample and Ri=i/N is the fractional rank of individual i in the SES distribution (i=1 and N for the poorest and wealthiest individuals, respectively). Given the relationship between covariance and ordinary least squares (OLS) regression, Kakwani48 showed that an equivalent estimate of the RC can be easily obtained from a ‘convenient regression’ as:
2where 
indexes the variance of fractional rank and the OLS estimate of β represents the RC.49 The SE of β provides an estimate of the SE of the RC.
The RC ranges from −1 to +1, with zero indicating perfect equality. A negative value of the RC index suggests that the health outcome is concentrated among poorer individuals and vice versa.50 Koolman and Van Doorslaer51 showed that if the estimated RC is multiplied by 75, it indicates the proportion of the health variable that would need to be redistributed (taken from the poorer half of the population and given to the wealthiest half if the health variable is concentrated among the poor) in order to achieve perfect equality.
As demonstrated by Wagstaff,52 the minimum and maximum values of the RC are not −1 and +1 and depend on µ when the health-related variable of interest is binary. In these cases, the index can be normalised by multiplying the β by 1/1−µ. Since the outcome variable in our analysis is binary, we used the normalised RC to quantify wealth-related inequalities in daily ESHS among children.
The RC quantifies relative differences in health outcomes across different SES groups. The (standard) concentration curve can be generalised in such a way that it becomes sensitive to variations in the mean of the health variable in the population, µ, and hence reflects absolute differences in health across socioeconomic groups. The absolute (generalised) concentration curve can be obtained if we multiply the standard concentration curve by µ. The absolute concentration curve demonstrates the cumulative percentage of the population, ranked according to an SES variable, against the cumulative amount of the health variable. The absolute (generalised) concentration index (AC) is defined as twice the area between the absolute concentration curve and the diagonal. The AC ranges from −µ to µ, with zero indicating ‘perfect equality’ and can be computed as:53
3
We applied sampling weights in the calculation of the RC and AC to obtain estimates that are representative of children (0–5 years) living in each country. Additionally, we accounted for clustering of children at the household level in the analysis. A 95% CI was calculated to assess the precision of our estimates and evaluate whether they were significantly different from zero (indicating perfect equality). A method proposed by Altman and Bland54 was employed to examine the significance of differences in RC and AC indices at the p value=0.05 level with 95% CIs. Moreover, we estimated measures of socioeconomic inequality in daily ESHS across countries by ranking countries based on their GDP per capita and then calculating the RC and AC. Furthermore, population-weighted Pearson correlation coefficients were used to assess the associations between country-level socioeconomic indicators (ie, log GDP per capita and adult (+15) literacy rate) and rates of adult smoking and daily ESHS across the sampled countries.
Decomposition of socioeconomic inequalities
To calculate the contributions of rural versus urban residence to within-country inequalities in ESHS, we decomposed the RC and AC for the countries where we observed statistically significant socioeconomic inequality. The RC index of daily ESHS in each country can be decomposed using the following formula:55 ,56
4where RC is the relative concentration of ESHS in the population as a whole. RCB represents relative inequality between groups (rural–urban) and is computed by assigning all individuals in each region the mean value of the ESHS in that region. The between-group inequality measures the impact (decrease) on the RC index if we were to redistribute children's ESHS in rural and urban areas such that both regions had the same rates of ESHS as the whole population. RCj is the RC of ESHS in region j and γj is the product of region j's share of ESHS and its share of the total population. The γ-weighted sum of RCj indicates the average degree of inequality within regions. The within-group inequality measures the impact (decrease) on the RC index if we were to redistribute children's ESHS within urban and rural regions such that wealthy and poor children in each region were equally exposed to secondhand smoke. R is a reranking term and is calculated as a residual.57 The reranking component is equal to zero if the subgroup SES ranges do not overlap. In other words, assuming that the total population is divided into two regions, rural and urban, where individuals living in urban areas are more socioeconomically well-off than those living in rural areas, R is equal to zero if the poorest person living in the urban region is better-off than the wealthiest person in the rural region.55 Using Wagstaff's correction to normalise the RC index yields:
5The decomposition of the AC can also be formulated as:
6
Results
Daily ESHS among children
Table 1 presents the sample sizes, GDP per capita, adult (+15) literacy rate, adult smoking rate, and total and rural-specific/urban-specific prevalence of daily ESHS for each country. As reported in table 1, 32% of children aged 0–5 years in the sampled countries were exposed to secondhand smoke on a daily basis at home (note: using data from the UN Department of Economic and Social Affairs, Population Division,58 we applied total number of children aged 0–5 years during the study period in each country58 as a weight in the calculation). In the majority of countries, >20% of children were exposed to secondhand smoke. There was a substantial variation across countries; for example, in Honduras, Tajikistan and Ethiopia <10% of children were exposed to secondhand smoke, whereas this figure was >60% in Indonesia, Jordan and Armenia. As illustrated in figure 1, the prevalence of daily ESHS was generally higher in Asia and the Middle East than in sub-Saharan Africa. The results also indicated that children's ESHS was higher in rural areas compared to urban areas in the majority of countries. While median daily ESHS in urban areas was 18% (IQR=19%) across countries, this figure was 24% (IQR=27%) among children in rural areas. The results also indicated statistically significant positive correlations between the prevalence of daily ESHS among children, and the (log) GDP per capita (r(24)=0.48, p=0.01) and adult (+15) literacy rate (r(24)=0.69, p<0.001), as shown in online supplementary figure A.1. The positive correlation was also found between the adult smoking prevalence for males, and the (log) GDP per capita (r(20)=0.68, p<0.001) and adult (+15) literacy rate (r(20)=0.82, p<0.001).
Survey year, sample size, GDP per capita and prevalences of adult smoking and children's daily ESHS in 26 LMICs
Prevalence of daily exposure to secondhand smoke among children aged 0–5 years across 26 low-income and middle-income countries.
Supplemental material
Socioeconomic inequality in daily ESHS
Table 2 contains estimates of the RC and AC for daily ESHS for 26 LMICs. Estimates of the RC and AC suggested that daily ESHS was concentrated among poorer children in almost all countries (RC, median=−0.179, IQR=0.186 and AC, median=−0.040, IQR=0.055). These results confirm our descriptive findings that indicated 36% of children in the lowest wealth quintile were exposed to daily ESHS, whereas the overall ESHS for children in the highest wealth quintile was 24%. Mali (RC=0.130; 95% CI 0.085 to 0.175 and AC=0.023; CI 0.015 to 0.031) and Tajikistan (RC=0.281; CI 0.203 to 0.360 and AC=0.018; CI 0.013 to 0.023) were the only two countries where daily ESHS was more concentrated among children living in wealthier households. There were substantial variations in the magnitudes of socioeconomic inequalities in daily ESHS across countries. As shown in figure 2A, relative socioeconomic inequalities in ESHS were highest in Benin and Gabon, whereas the highest absolute socioeconomic inequalities were observed in Indonesia and Cambodia (figure 2B). Although a statistically significant positive correlation was observed between the prevalence of daily ESHS among children and (log) GDP per capita, we did not observe any relation between socioeconomic inequalities in daily ESHS and (log) GDP per capita.
Relative and absolute concentration indices for daily ESHS among children aged 0–5 years in 26 LMICs
Log gross domestic product per capita versus relative and absolute inequalities in daily exposure to secondhand smoke among children aged 0–5 in low-income and middle-income countries.
Turning to rural-specific/urban-specific estimates, we found that socioeconomic inequalities in daily ESHS were concentrated among children belonging to lower SES households in rural and urban areas. Tajikistan was the only country where the RC and AC suggested that daily ESHS was more concentrated among children living in wealthier households in urban (RC=0.092; CI 0.003 to 0.181 and AC=0.01; CI 0.001 to 0.019) and rural areas (RC=0.044; CI 0.01 to 0.079 and AC=0.003; CI 0.001 to 0.005). Daily ESHS was also more concentrated among children of wealthier households in rural areas in Mali (RC=0.067; CI 0.01 to 0.124 and AC=0.011; CI 0.002 to 0.02). In addition, the concentration of ESHS among socioeconomically disadvantaged children was greater in magnitude in urban relative to rural areas. This latter finding is consistent with our descriptive statistics that showed greater differences in the prevalence of daily ESHS (%) between children in the lowest and highest wealth quintiles in urban areas (15=33–18) compared to rural areas (7=38–31).
Summary measures of socioeconomic inequality in ESHS, calculated by ranking countries based on their GDP per capita, demonstrated that daily ESHS was concentrated among wealthier countries (RC=0.281; CI 0.082 to 0.48 and AC=0.089: CI 0.026 to 0.152). These results confirm the statistically significant positive correlations that we observed between the prevalence of daily ESHS among children and log GDP per capita across countries.
Decomposition of socioeconomic inequality in daily ESHS
Table 3 and online supplementary figure A.2 reported the results of the decomposition of socioeconomic inequality in daily ESHS for each country into between-group and within-group components. The reported contributions of between-group inequality in Benin, for example, suggest that if we were to redistribute children's ESHS in rural and urban areas such that both regions had the same prevalence of daily ESHS as the whole population (ie, 0.10), the RC and AC indices of daily ESHS would be reduced by the absolute values of 0.104 (ie, from −0.425 to −0.311) and 0.011 (ie, from −0.044 to −0.033), respectively. The contributions of within-group inequality in Benin indicate that if we were to redistribute children's daily ESHS within urban and rural areas such that poor and wealthy children in each region were exposed equally to secondhand smoke, the RC and AC indices of daily ESHS in Benin would be reduced by the absolute values of 0.183 (ie, from −0.425 to −0.242) and 0.019 (ie, from −0.044 to −0.025), respectively.
Decomposition of relative and absolute concentration indices for daily ESHS among children aged 0–5 years into between- and within-group components in LMICs
The decomposition of RC and AC indices demonstrated that inequality of daily ESHS within rural and urban areas was the main source of socioeconomic inequality at the national level in most countries (median=46%, IQR=32%). The contribution of within-group (rural and urban) inequality to the overall socioeconomic inequality was pronounced in countries such as Rwanda, Burundi, Jordan, Nepal, Uganda, Bangladesh, Cambodia and Tajikistan (table 3 and online supplementary figure A.2). The results indicated that children in rural areas in all of the countries, except Mali and Tajikistan, are systematically exposed to more daily secondhand smoke at home compared to those in urban areas (as shown by the negative values of the between-group component in table 3). The median contribution of inequality of daily ESHS between rural and urban areas to the overall measures of socioeconomic inequality was 27% (IQR=41%). The between-area inequality in daily ESHS contributed substantially to overall socioeconomic inequality in Congo Brazzaville, Pakistan, Mali, Zimbabwe and Kyrgyz Republic.
Discussions
Passive smoking is generally more common among children growing up in socioeconomically deprived households and may contribute to social inequalities in health experienced over the life course. Notwithstanding the importance of ESHS, few studies have examined socioeconomic inequalities in ESHS among children in poorer countries. Using information collected through the DHS, we quantified and decomposed socioeconomic inequalities in children's ESHS at home in 26 LMICs.
Our findings indicated a large variation across LMICs in the prevalence of daily ESHS among children aged 0–5 years. The prevalence of daily ESHS was higher in countries with greater per capita income and literacy rates. Similar to results from higher income countries,25–29 ,59 we observed a negative socioeconomic gradient in all but two (ie, Mali and Tajikistan) of the 26 LMICs. Children's ESHS was higher in rural compared to urban areas in the majority of countries. In contrast, the concentration of ESHS among socioeconomically disadvantaged children was greater in urban compared to rural areas.
Although the difference in ESHS between rural and urban areas contributed to the overall socioeconomic inequality in ESHS among children in LMICs, our findings demonstrated that socioeconomic inequality within rural and urban areas was the main driver of the overall socioeconomic inequality in most countries. These results suggest that strategies to reduce inequalities in ESHS among children at home should address the prevalence of passive smoking among poorer children in urban and rural areas in developing countries.
Population-level tobacco control strategies include interventions to: increase tobacco prices, such as cigarette taxes; ban smoking in public venues, including workplaces, restaurants and bars; improve public knowledge of the harms of tobacco use through educational campaigns and support smoking cessation, among others.60 There is growing evidence that population-level strategies reduce the prevalence of tobacco use and ESHS. For example, results from a systematic review suggest that smoke-free policies are effective at reducing the prevalence of smoking and encouraging cessation.61 Similarly, smoke-free public place policies appear to be effective at reducing secondhand smoke exposure in a variety of public settings, including workplaces, restaurants and bars.62–65 Given the association between being employed in a smoke-free environment and living in a smoke-free home, even in LMICs,66 ,67 tobacco control programmes that prohibit smoking in public places might shift social norms regulating smoking behaviour and encourage smoke-free policies at home, as observed in the USA, Australia and New Zealand.60 For example, a cross-sectional analysis suggested that legislation banning smoking in public places in Scotland was associated with considerable reductions in ESHS among non-smoking adults, and stimulated the adoption of smoke-free homes and cars.68
Despite reliable evidence concerning the effects on smoking prevalence of population-level tobacco control strategies, particularly bans, implications for health inequalities are less clear. Several reviews69 ,70 have examined impacts on inequalities in smoking prevalence among adults. In general, socioeconomically disadvantage groups are more responsive to cigarette price/tax increases and these policies have been consistently associated with larger reductions in smoking among lower compared to higher SES groups. In contrast, the effects of smoke-free policies on social inequalities in smoking are mixed; comprehensive bans appear to benefit socioeconomic groups equally, but voluntary or partial policies have the potential to exacerbate social inequalities in smoking prevalence. The effect of media and educational campaigns, as well as cessation services, on social inequalities is equivocal and may depend on whether these interventions are targeted to specific populations. Few studies have explored the equity dimensions of tobacco control interventions for young people. There is some evidence that cigarette price/tax increases have the potential to diminish social inequalities in smoking among youth,71 whereas the relation between the strength of tobacco control programmes and smoking among adolescents in European countries did not vary by household SES.72
Further work is needed to clarify if population-level tobacco control strategies are a viable lever for reducing social inequalities in ESHS, especially in LMICs. Recently, Moore et al73 found that legislation disallowing smoking in public places did not appear to influence social inequalities in ESHS in Wales. A study from Taiwan,62 however, indicated that the expansion of smoke-free legislation reduced inequalities in ESHS across educational but not income groups. Both studies compared ESHS before and shortly after these reforms and lacked control groups that were unaffected by the policies; it was therefore difficult to examine longer term effects or account for secular trends in ESHS.
There are limitations to our study. First, self-reported measure of daily ESHS could be an issue in our analysis if under-reporting was associated with SES. For example, if poorer women were less likely to report smoking in their homes due to conservative environments in lower SES households, our study may underestimate the concentration of ESHS among poorer children.74 Second, as our variable of interest in the study is bounded, the minimum and maximum values of the RC (AC) are not −µ (−1) and µ (1) and, as Wagstaff52 demonstrated, depend on the µ.75 There was much lively debate as to how to overcome this issue in health economics literature.76–78 Whereas Wagstaff suggests multiplying the concentration index by 1/1−µ, Erreygers suggests multiplying the concentration index by 4µ75 ,79–81 when the outcome variable is binary. In recent work, Kjellsson and Gerdtham76 argue that while Wagstaff's index52 shows the extent to which society is far from a state where individuals at the top of the income/wealth distribution are healthy (ie, exposed to secondhand smoke), the Erreygers' index75 demonstrates the extent to which the society is far from a state where the upper 50% of individuals in the wealth/income distribution are healthy, irrespective of µ. Since we used Wagstaff correction to overcome the binary nature of daily ESHS variable, our results can be interpreted according to the value judgment inherent in the Wagstaff index. We also performed sensitivity analyses using Erreygers' correction in the calculation of the RC and AC. As reported in online supplementary table A.1, the calculated RC and AC informed qualitatively similar inference.
Conclusion
Our study indicated distinct variation in children's exposure to indoor smoking levels among LMICs. Children from rural and socioeconomically disadvantaged households (particularly those living in urban areas) are more likely to experience daily ESHS at home. Our results suggest that special attention should be devoted to address ESHS among these groups because socioeconomic disparities in ESHS might lead to socioeconomic disparities in tobacco-related health problems over the life course. The inequalities in ESHS between poorer compared to wealthier children should be seen as a health inequity because they are preventable and unfair.82 Additional research is needed to identify mechanisms for reducing social inequalities in secondhand smoke exposure among children in LMICs.
What this paper adds
Children's daily exposure to indoor smoking was higher in rural areas compared to urban areas in the majority of the countries.
The concentration of ESHS among poorer children was greater in urban relative to rural areas.
Wealth-based differences in ESHS within urban and rural areas were the main contributor to socioeconomic inequalities in most countries.
Special attention should be given to reduce ESHS among children from rural and socioeconomically disadvantaged households.
References
Footnotes
Contributors MH and AN contributed to the conception and design of the study, MH performed the statistical analysis and drafted the manuscript and AN helped with drafting and revisions. MH and AN read and approved the final version of the manuscript.
Funding MH acknowledges funding for this research provided by the Canadian Institutes of Health Research (CIHR) fellowship award program. AN acknowledges the support of the Canada Research Chairs Program. Both authors acknowledge funding from the Canadian Institutes of Health Research Operating Grant, ‘Examining the impact of social policies on health equity’ (ROH-115209).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement All data underlying the findings are fully available on request from the Demographic Health Survey (DHS) program.