Healthy People 2020: Social Determinants of Cigarette Smoking and Electronic Cigarette Smoking among Youth in the United States 2010–2018

The purpose of this study was to determine social determinants of cigarette smoking and ever using electronic cigarettes (e-cigarettes) among young adults aged 18 to 25 years in the United States between 2010 and 2018. Using secondary data from National Health Interview Surveys (NHIS) across the 2010, 2014, and 2018 survey years, this study analyzed the prevalence rates of cigarette smoking and ever using e-cigarettes between 2010 and 2018, demographic and socioeconomic disparities in smoking, and the relationship between previous e-cigarette use and current smoking. First, the past decade witnessed a notable decline in conventional cigarette smoking and a sharp increase in e-cigarette use among youth. These trends were consistent regardless of socioeconomic status. Second, demographic and socioeconomic disparities persisted in cigarette smoking. Non-Hispanic white male youth were more likely to become smokers as they grew older. Young people with lower educational attainment, living below the U.S. federal poverty level, and having a poor physical health status had a higher smoking prevalence. Third, previous e-cigarette use was more likely to relate to subsequent cigarette use among young people. To achieve the Healthy People 2020 objectives, tobacco control programs and interventions need to be more specific in higher prevalence groups and service providers should not assume that there is a one-size-fits-all model for youth.


Introduction
By the end of 2017, the United States had over 34 million current cigarette smokers, 14% of U.S. adults, who smoked either every day or some days during the typical week, a significant decline from 20.9% in 2005 [1]. This prevalence rate was ten percent among young adults aged 18 to 24 years. Despite the falling prevalence of smoking, cigarette smoking still has significant public health consequences. In a report which summarized 50 years of progress in tobacco control and prevention, the Office of the Surgeon General [2] specified that more than 16 million Americans suffered from a disease caused by smoking. Many diseases, 87 percent of lung cancer deaths, 32 percent of coronary heart disease deaths, and 79 percent of all cases of chronic obstructive pulmonary disease, were caused by smoking, and smoking has become the leading cause of preventable disease and death (about 1 in 5 deaths) in the United States [3].
The large prevalence of tobacco use has become one of the twelve leading health indicators put forward by Healthy People 2020, which was led by the U.S. Surgeon General and provided a national agenda for improving the health of the nation and achieving health equity for the period

Study Design, Data and Sample
This study utilized the secondary data from the National Health Interview Survey (NHIS) across three waves to study the trend and prevalence of cigarette smoking and electronic cigarette smoking as well as social determinants of smoking among young adults in the past decade. Since 1957, the U.S. Census Bureau has been the data collection agency for the NHIS and it uses household-based sampling to collect data on an annual basis among the non-institutionalized U.S. civilian population nationwide. The data mainly collected are information on health status, health care access, and progress which are frequently used as major data sources to evaluate national health objectives [12]. Because the year 2018 was the latest dataset available for public use, this study utilized data from these three waves, 2010, 2014, and 2018, to represent the health trend in the past decade. Out of these three survey waves, data in 2010 represent the beginning of the decade, data in 2014 represent the middle point of the decade, and data in 2018 represent the end point of the decade. The inclusion criterion had resulted in samples which only included young adults aged 18 to 25 years old. For each wave of data, the NHIS included several data files such as person file, sample adult file, and family file. Using the unique person identifier, the authors merged person, sample adult, and family files into one combined dataset for each wave. The sample sizes for three waves were 3281 respondents in 2010, 3981 in 2014, and 2195 in 2018, respectively.
The NHIS datasets are available for download from the Centers for Disease Control and Prevention (CDC) website for public use. All the downloadable public use data files are not individually identifiable, and there is no way to link data with the subjects from whom the data were originally collected. Therefore, this study did not involve human subjects and was not subject to approval from the Institutional Review Board (IRB) office.

Measurements
Smokers were identified using two major variables. Current cigarette smokers were those who reported having smoked every day or some days of the typical week at the time of survey. The indicator for current smoker was measured as 1 if the respondent smoked every day or some days, and 0 if not smoking at all. Ever electronic cigarette smokers were those who ever smoked electronic cigarettes at the time of survey. The indicator for an ever e-cigarette smoker was measured as 1 if the respondent ever used electronic cigarettes and 0 if never. Although current cigarette smokers were measured each wave, ever e-cigarette smokers were only measured in recent waves 2014 and 2018. Former smokers were excluded from analyses of smoking status because there were overlaps between current smoking and former smoking status.
Several sociodemographic variables were measured in the study of disparities in smoking prevalence. Demographic variables included age, gender, and race/ethnicity. Age was continuous in actual years old. Gender was dichotomous with 1 = male and 2 = female. Race/ethnicity was coded as four groups: 1 = Non-Hispanic white, 2 = Non-Hispanic black, 3 = Hispanic, and 4 = other. Socioeconomic variables included education, income-to-poverty ratio, self-reported health status, and citizenship status. Education and income are considered as two most important indicators of socioeconomic status. Based on the original variable, "R's highest schooling", education was recoded into four categories: 1 = less than high school, 2 = high school or General Education Development (GED) equivalent, 3 = some college or associate degrees, 4 = bachelor and above. The income-to-poverty ratio was created to proxy the family's financial security and the poverty threshold took into account both income and family size. In general, a ratio less than one implied that the income was less than the set poverty threshold and those people whose incomes fell under the threshold were considered poor and in poverty. The income-to-poverty ratio was measured in three categories: 1 = less than 100% U.S poverty level, 2 = between 100% and 199%, and 3 = 200% and above. The self-reported health status was a categorical variable, where the categories had an intrinsic order with 1 = excellent, 2 = very good, 3 = good, and 4 = fair/or poor. The citizenship indicated whether respondents had U.S. citizenship with 1 = yes and 0 = no.

Analyses
Descriptive statistics were reported for all variables across three survey waves. In addition to frequencies and proportions, the line charts were drawn to represent the trends over time. Logistic regressions were conducted to test the relationship between sociodemographic factors, being a previous electronic cigarette smoker, and being a current smoker. Odds ratios and 95% confidence intervals were reported. All the statistical analyses were completed by using SPSS version 24.0, IBM North America, New York, United States.   Table 2 presents the prevalence rates of current cigarette smokers and electronic cigarette smokers among youth during the period 2010-2018. These proportions were broken down by age, gender, race/ethnicity, education, income-to-poverty ratio, reported health status, and citizenship status. Figures 1-5 depict these trends in terms of line graphs. The trends were consistent across different sociodemographic indicators. Specifically speaking, the prevalence rates of cigarette smoking declined in the past decade with the prevalence rate highest in 2010 and lowest in 2018. In a sharp contrast, electronic cigarette smokers increased from 2014 to 2018.          Table 3 shows the results from logistic regression predicting the likelihood of being a current cigarette smoker by controlling sociodemographic variables and previous electronic cigarette smoking status. Each column reports respective results from each wave. As mentioned above, electronic cigarette data were not collected in 2010, therefore the ever e-cigarette indicators were only included in the analyses for two survey years, 2014 and 2018. In both waves, after controlling sociodemographic factors, previous electronic cigarette smokers in 2014 (p < 0.001, OR = 10.428, 95% CI 8.502-12.790) and 2018 (p < 0.001, OR = 6.666, 95% CI 4.770-9.316) were more likely to be current smokers, compared to non-smokers. Demographic characteristics also strongly predicted the likelihood of being a current cigarette smoker versus a non-smoker. As youth grew older, they were more likely to become smokers and the results were statistically significant across three waves in 2010 (p < 0.001, OR = 1.193, 95%      Table 3 shows the results from logistic regression predicting the likelihood of being a current cigarette smoker by controlling sociodemographic variables and previous electronic cigarette smoking status. Each column reports respective results from each wave. As mentioned above, electronic cigarette data were not collected in 2010, therefore the ever e-cigarette indicators were only included in the analyses for two survey years, 2014 and 2018. In both waves, after controlling sociodemographic factors, previous electronic cigarette smokers in 2014 (p < 0.001, OR = 10.428, 95% CI 8.502-12.790) and 2018 (p < 0.001, OR = 6.666, 95% CI 4.770-9.316) were more likely to be current  In addition, the prevalence of smoking in youth varied widely in terms of their socioeconomic status. Compared to youth who had bachelor or graduate degrees, youth without four-year college education were significantly more likely to smoke cigarettes and the lower the education, the higher likelihood of smoking. In regard to income-to-poverty ratio, lower income was significantly associated with higher likelihood of smoking in 2010 and 2018. Youth whose families were living below the 100% poverty line had the highest likelihood of being current smokers (p = 0.008, OR = 1.363, 95% CI 1.083-1.716) in 2010 and (p = 0.016, OR = 1.621, 95% CI 1.095-2.400) in 2018. Youth whose families lived between the 100% and 200% poverty lines also had a higher likelihood of being current smokers (p = 0.005, OR = 1.428, 95% CI 1.115-1.829) in 2010, compared to youth whose family incomes were above 200% income-to-poverty ratio. The likelihood of smoking also varied widely according to youth's self-reported physical health status. Compared to those who reported excellent health, youth whose health statuses were very good/good/fair/poor were more likely to become current smokers and the fair/poor sample had the highest likelihood of smoking (p < 0.001, OR = 2.883, 95% CI 1.943-4.277) in 2010. Youth who were U.S. citizens were significantly more likely to smoke cigarettes than immigrant youth in 2010 (p < 0.001, OR = 2.135, 95% CI 1.425-3.199).

Discussion
This study examined a population-based national sample of U.S. young adults whose ages were between 18 and 25 years old, in three survey years from 2010 to 2018. Research findings reported that prevalence rates of smoking behaviors varied widely across sociodemographic characteristics and there were strong and significant relationships between previous electronic cigarette smoking and current smoking status. The proportion of smokers in the US had significantly decreased during the period 2010 to 2018, and this trend was consistent regardless of age, gender, race/ethnicity, education, income-to-poverty ratio, self-reported health status, and citizenship. This declining trend among U.S. youth is similar to findings by the CDC [13], stating that only ten percent of young adults aged 18 to 24 years smoked cigarettes in 2017. This falling trend is also similar to that of other countries such as Argentina and Brazil, to name a few [14,15]. Youth were more likely to become cigarette smokers as they grew older. There were more male smokers than female smokers. Non-Hispanic white youth were significantly more likely to be smokers compared to other racial/ethnic youth. Socioeconomic status played an important role in predicting the likelihood of being a smoker. Citizens were associated with a higher prevalence of cigarette smoking compared to immigrant youth. Youth who had lower levels of educational attainment, who lived below or near the U.S. federal poverty levels, and who had poor physical health status were very likely to become current smokers. These findings were consistent with previous research on cigarette smoking among people of low socioeconomic status [16,17]. Studies by Carlson et al. (2018) identified barriers to quitting smoking among youth with a low socioeconomic status and concern about weight gain was statistically associated with quitting status. These findings also suggested that risks associated with cigarette smoking may be evident as these social determinants clearly resulted in negative health outcomes.
There are important implications based on the findings of this study. First, comprehensive tobacco control policies, anti-smoking campaigns and educational programs may have contributed to the notable decline in cigarette smoking among young people in the past decade. Since 1965, the U.S. Surgeon General has been generating annual reports on smoking and health which inform the general public of the health consequences of smoking and changes in the tobacco landscape. More than half of the fifty U.S. states and a growing number of cities and counties have enacted strong smoke-free laws that require workplaces and public places such as restaurants and bars to be smoke-free [18]. These tight government regulations successfully created barriers for smokers and reduced the prevalence of smoking. In addition, most commercial health insurance plans and Medicaid insurance were required to provide coverage on smoking cessation programs to help smokers to quit smoking [19]. In addition to these nationwide laws and initiatives, numerous anti-smoking campaigns and educational and prevention programs gave young people tools and resources to prevent smoking at early ages. Second, while the notable decline in smoking rates is a public health success, the socioeconomic inequality in cigarette smoking prevalence rates is still astonishing. In order to narrow the health disparity gap, tobacco control initiatives need to be more specific by targeting groups with higher smoking prevalence. For youth whose families are of low socioeconomic status, social service and health care providers might provide culturally competent smoking cessation programs for different racial/ethnic youth because they have different cultures and environments as regards cigarette smoking. Providing culturally competent programs may become an effective approach in smoking cessation and prevention. As smoking is a learned and socially-mediated behavior [20], youths are heavily influenced by social norms in their immediate circle and are very likely to experiment with tobacco use if family members or friends smoke. Third, in sharp contrast to the decline in conventional cigarette smoking, use of electronic smoking is dramatically rising among youth. Although unsafe health effects of e-cigarette use among U.S. youth and young adults are clearly documented by the CDC [21], there are no studies on the long-term effects of e-cigarette use due to the fact that e-cigarettes are new nicotine-delivery products on the market. The investigation of lung injury associated with use of e-cigarettes will continue. Fourth, since Healthy People 1990 was first made public by the U.S. Surgeon General in 1979, Healthy People initiatives have been consistently providing overarching health objectives and targets for four decades (Healthy People 1990, 2000, 2010, and 2020). While each decade was unique, achieving health equity and eliminating health disparities were consistent and coherent health objectives across four decades. As indicated by findings from this study, cigarette prevalence rates are still unevenly distributed and widely varied across demographic and socioeconomic status. Out of the five key areas of social determinants of health, this study analyzed three areas, namely, education, economic stability, health and health care, and did not touch the other two areas, social and community context, neighborhood and built environment. Therefore, future research has to continue exploring other key areas of social determinants of smoking and identifying effective approaches to prevent tobacco use.
There are several limitations of this study that deserve additional attention. First, this study used a repeated cross-sectional design by examining changes in trends over time in the past decade. Unlike a longitudinal cohort study which represents changes within the same individuals, this cross-sectional study is a snapshot and prevents examination of the longitudinal association between social determinants and current smoking behaviors. Thus, this study cannot establish a causal relationship between social determinants and smoking prevalence. Second, this study utilized data from three survey years-2010, 2014, and 2018, to represent the past decade. Once the 2019 wave data are available, researchers will be able to extend this study by using data from 2010-2019 to represent the time trend during the past decade. With that said, 2010 represents the beginning of this decade, 2015 is the middle point of the decade, and 2019 is the end of the decade. Third, data were limited to youth aged 18-25 and therefore cannot be generalized to other age groups.
Despite methodological limitations, this study significantly contributed to bridging the gap in knowledge about trends in smoking over the last decade, demographic and socioeconomic disparities in smoking, and associations between previous electronic cigarette usage and current smoking status among U.S. young people aged 18 to 25 years. Using comprehensive and representative national-level survey data, the results of this study can shed light on the social determinants of smoking prevalence among young people worldwide.

Conclusions
This was the first study to examine the social determinants of cigarette smoking prevalence and associations between previous electronic cigarette usage and current cigarette smoking among youth aged 18 to 25 in nationally-representative U.S. samples during the period 2010 to 2018. Although overall smoking prevalence has decreased in the U.S. over the past decade (2010-2018), socio-economic disparities related to cigarette smoking still persist. Young people with lower educational attainment, living below the U.S. federal poverty level, and having poor physical health status had a higher smoking prevalence. Youth who previously smoked electronic cigarettes were more likely to become current smokers. To achieve the Healthy People objectives of achieving health equity and eliminating health disparities, tobacco control programs and interventions need to be more specific and service providers should not assume that there is a one-size-fits-all model for youth with dramatically different demographic and socioeconomic characteristics. Interventions need to be sensitive and specific in higher prevalence groups such as youth with lower social economic status. There is still room for improvement. As young adulthood represents a distinct developmental period of the life course, early intervention and prevention programs for smoking cessation will significantly reduce adverse effects and benefit society as a whole. Moreover, public health policies, regulations, and prevention programs need to prioritize strategies which address the shifting trends toward electronic cigarette usage among young adults.
Author Contributions: G.W. led this study and drafted the manuscript. L.W. conducted analyses and contributed to the writing. All authors have read and agreed to the published version of the manuscript.

Funding:
We gratefully acknowledge the support of the Research Project of Young Scholars, "Innovations of Data Protection Regulations", in Humanities and Social Sciences of Wuhan University.

Conflicts of Interest:
The authors declare no conflict of interest.