Smoking and Health-Related Quality of Life Among U.S. Adolescents

Abstract

Objective:

Smoking continues to be a public health problem among youth. Developmentally, adolescence is a period marked by the vulnerability to initiate risk behaviors such as smoking. While studies have documented associations between smoking and poor health related quality of life (HRQOL) among adults, little is known about the association among adolescents.

Methods:

Data on smoking and HRQOL from a sample of 4,848 adolescents aged 12–17 years from the 2001–2008 National Health and Nutrition Examination Surveys were analyzed. Smoking status (current, not current, and never) was determined using self-report data and serum cotinine levels. HRQOL was assessed based on self-reported physical and mental health in the last 30 days, activity limitations in the last 30 days, and general self-rated health.

Results:

Compared with never smokers, adolescents who ever smoked reported more recent physically unhealthy days (p < .001), mentally unhealthy days (p < .0001), and activity limitation days (p < .01). Compared with never smokers, adolescents who ever smoked or who were current smokers were more likely to report ≥14 physically unhealthy days, ≥14 mentally unhealthy, ≥14 activity limitation days, and fair or poor health; not current smokers were also more likely than never smokers to report ≥14 days for being both physically unhealthy and mentally unhealthy.

Conclusions:

Among a nationally representative sample of adolescents, this study found strong associations between smoking and HRQOL measures. The relationship of smoking to self-reported activity limitations warrants attention and further research. The findings underscore the importance of addressing smoking and subjective well-being early in the lifespan.

The original version was incorrect. The reference to DiFranza et al. (2007) has been corrected.

Introduction

Cigarette smoking continues to be the leading cause of morbidity and mortality in the United States and worldwide (United States Department of Health and Human Services, 2004). Despite significant progress, smoking among youth remains a serious public health problem, especially since 80% of adult smokers begin smoking before the age of 18 (Substance Abuse and Mental Health Administration, 2009)). For example, in the United States, 1.5 million youth aged 12–17 years began smoking cigarettes in 2009, at least 4,000 per day (Substance Abuse and Mental Health Administration, 2009). In 2009, close to one in two high school students (46.3%) reported ever trying cigarettes, and one in five (19.5%) reported current smoking (Centers for Disease Control and Prevention, 2010c). While youth smoking trends sharply declined from 1997 to 2003, the declines in recent years have slowed (Centers for Disease Control and Prevention, 2010a, 2010b, 2010c), likely resulting from decreases in tobacco control funding, fewer cigarette price increases, and a continuation of pro-tobacco influences. These data underscore that even with effective policies, regulations, and programs, smoking among youth is still pervasive.

Adolescence can be a stressful and vulnerable period of development, marked by a tendency to experience poor mental health, which may increase the propensity to adopt risky behaviors such as tobacco use (Schulenberg, Maggs, & Hurrelmann, 1999). For example, stress, psychological distress, trait anxiety, and depressed mood are associated with early onset smoking and continued smoking into adulthood (Anda et al., 1999; Degenhardt & Hall, 2001; DiFranza et al., 2004; Dube, Cook, & Edwards, 2010; Dube, Felitti, Dong, Giles, & Anda, 2003; Jasuja, Chou, Riggs, & Pentz, 2008; Jorm et al, 1999; Jun et al., 2008; Lawrence Mitrou, Sawyer, & Zubrick, 2010; Tercyak, Goldman, Smith, & Audrain, 2002; van Loon, Tijhuis, Surtees, & Ormel, 2005). In fact, a prospective study found that youth who experienced relaxation after first dose of nicotine and youth experiencing depressed mood at the first dose of nicotine were more likely to progress to smoking (DiFranza et al., 2007).

Measures of self-reported subjective health appraisals can assess individuals’ experience, internal processing, and feeling, as it relates to multiple dimensions of health and well being: physical, social, emotional, mental, and intellectual (Kindig, Asada, & Booske, 2008). Subjective assessments of health among adolescents are critical to measure, as they provide a perspective of well being not otherwise captured. Health-related quality of life (HQROL) is a broad multidimensional health outcome that provides information beyond more traditional indicators of health such as mortality and morbidity (Kindig et al., 2008) and usually includes self-reported measures of physical and mental health (Ferrans, 2005; Moriarty, Kobau, Zack, & Zahran, 2005).

While associations between poor HRQOL and psychological distress have been reported among current smoking adults in the United States and in other adult populations (Dube et al., 2009; Heikkinen, Jallinoja, Saarni, & Patja, 2008; McClave et al., 2009; Schmitz, Kruse, & Kugler, 2007), very few studies have assessed these relationships among adolescents. In a study using data from the South Carolina Youth Risk Behavior System, youth smoking and the total number of recent unhealthy days were found to be significantly associated (Zullig, Valois, Huebner, & Drane, 2005); however, general self-rated health and activity limitations were not included in these analyses. Based on findings from the U.S. Health Behavior in School-Aged Children Study, adolescents who smoked daily or who experimented with cigarettes reported more subjective somatic and psychological health complaints than those who did not smoke (Botello-Harbaum, Haynie, Murray, & Iannotti, 2011); however, specific HRQOL measures were not used in this study.

To provide further evidence of the relationship between adolescent smoking and subjective well being, we sought to assess the association between current smoking and HRQOL among adolescents aged 12–17 years, using data from the 2001–2008 National Health and Nutrition Examination Survey (NHANES), which includes questions regarding physically and mentally unhealthy days, activity limitations, and general health as well as self-reported smoking history and serum cotinine. We examine if there are differences in physically and mentally unhealthy days, activity limitation days, and self-rated health between youth who ever smoked and youth who never smoked; we also examined differences using a three-group status: current smokers, not current smokers, and never smokers.

Methods

Sample

NHANES is a nationally representative multistage survey of noninstitutionalized U.S. civilians of all ages conducted by the National Center for Health Statistics and released in two-year cycles. The study design includes face-to-face interviews and a medical examination component (MEC). The sample design and weighting methodology for NHANES have been described previously (Centers for Disease Control and Prevention, 2011). Briefly, primary sampling units were single counties or multiple, less populous counties combined to meet a minimum population size requirement. Clusters of households were selected from these counties, and each person in a selected household was screened for demographic characteristics. After initial screening, one or more individuals from a household were selected for the sample.

A total of 6,108 adolescents aged 12–17 years participated in one of the four 2-year NHANES cycles from 2001–2002 through 2007–2008, of whom 4,848 (79.4%) had complete data available about their smoking behaviors, demographic characteristics, and HRQOL outcomes. Most of the missing data were variables related to smoking (16.5%) and the poverty-to-income ratio (5.6%). Participants with and without available data on either the smoking variables or the poverty-to-income ratio did not statistically differ on the HRQOL outcomes.

Serum Cotinine

Since 1999, serum cotinine measurements were available in NHANES for participants aged 3 years and older. Whole blood samples are collected by venipuncture, and serum cotinine is analyzed using an isotope-dilution liquid chromatography tandem mass spectrometry method. Further information regarding laboratory analysis of serum cotinine and quality control methods have been published in detail elsewhere (J. T. Bernert, Jr., et al., 1997; J. T. Bernert, McGuffey, Morrison, & Pirkle, 2000; Pirkle, Bernert, Caudill, Sosnoff, & Pechacek, 2006). Serum cotinine levels >10.0ng/mL are associated with active smoking within the past few days (J. T. Bernert et al., 2000).

Classification of Smoking Status

Of the 4,848 participants with complete data, responses to questions regarding current and past smoking behavior as well as serum cotinine levels helped categorize individuals into three mutually exclusive groups: never smokers, not current smokers, and current smokers. Never smokers reported never trying a cigarette and had serum cotinine levels ≤10.0ng/mL. Not current smokers reported smoking at least one cigarette but not in the last 30 days and had serum cotinine levels ≤10.0ng/mL. Current smokers either reported smoking in the last 30 days or had serum cotinine levels >10ng/mL, indicating recent smoking (Pirkle et al., 2006). Ever smoked group included not current smokers or current smokers. We carried out analyses using both two-group (never smoked versus ever smoked) and three-group comparisons (never, not current, current). We also analyzed the association of cotinine alone with HRQOL, but the interpretation of these analyses is debatable because cotinine levels will tend to misclassify youth who may have smoked intermittently in the past month.

Health-Related Quality of Life Measures

The CDC monitors HRQOL in the U.S. population through the Behavioral Risk Factor Surveillance System (BRFSS) and the NHANES. CDC has developed four core HRQOL measures, which have undergone cognitive testing, and have been shown to demonstrate content validity, construct validity, criterion validity, with the Rand Medical Outcomes Study Short-Form 36 health survey, predictive validity, test–retest reliability, and internal consistency (Centers for Disease Control and Prevention, 2000; Moriarty et al., 2005; Moriarty, Zack, & Kobau, 2003). The measures have also been evaluated among adolescents and shown to be valid in this age group (Zullig, Valois, Huebner, & Drane, 2004). The four measures include items that assess general self-rated health, the number of physically and mentally unhealthy days in the last 30 days, and the number of activity limitation days due to being physically or mentally unhealthy in the last 30 days. The self-rated health question asks “Would you say that in general your health is excellent, very good, good, fair, or poor?”The physically unhealthy days question asks “Now thinking about your physical health, which includes physical illness and injuries, for how many days during the past 30 days was your physical health not good?” The mentally unhealthy days question asks “Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?” The activity limitation days question asks “During the past 30 days, for about how many days did poor physical or mental health keep you from doing your usual activities, such as self-care, work, or recreation?”

For logistic regression analysis, general self-rated health was dichotomized into fair or poor health vs. good, very good, or excellent health (the reference group). Physically unhealthy days, mentally unhealthy days, and activity limitation days were dichotomized at 14 or more days vs. 13 or fewer days in the past 30 days (≥14 days were classified as frequent physical distress, frequent mental distress, and frequent activity limitations). The ≥14-day criterion was used because clinicians and clinical researchers often use this time frame to assess the severity of clinical depression and anxiety disorders. Moreover, physical and mental health symptoms associated with longer durations tend to be associated with more activity limitations (Jiang & Hesser, 2006). These thresholds have also been utilized frequently in previous studies using large representative databases (Zahran et al., 2005; Heo, Allison, Faith, Zhu, & Fontaine, 2003; Jiang & Hesser, 2006).

Demographic Variables

Potentially confounding demographic variables available for statistical analyses and related to both smoking and HRQOL included: the NHANES cycle, age at interview, sex, race and ethnicity (non-Hispanic White, non-Hispanic Black, Mexican American, other Hispanic, other), and the poverty-to-income ratio (PIR). The PIR was calculated from self-reports by an adult NHANES participant regarding the household income and the number of individuals living in the household. The PIR represents the ratio of household income to federally defined poverty thresholds for the number of individuals in the household. A ratio below 1.00 indicates that the income for the respective household is below the officially defined poverty threshold, while a ratio at or above 1.00 indicates the household income is at or above this threshold.

Statistical Analyses

Weighted descriptive and regression analyses were carried out using SAS-callable SUDAAN 9.1.3 (Research Triangle Institute, Research Triangle Park, NC) to take into account the complex survey design of NHANES. Linear and logistic regression analyses were adjusted for the NHANES cycle, age at interview, gender, race and ethnicity, and the poverty-to-income ratio. In addition, we tested for effect modification of the smoking status variables by gender and race and ethnicity. We also tested whether the serum cotinine levels associations with the HRQOL outcomes were similar to the results found with the smoking status variables in order to provide supporting evidence and validation of the associations found with the self-report data. Because age and the highest school grade completed were highly correlated among adolescent participants (r = .92), only age was included in the statistical analyses. Associated p values were presented for all comparisons. Predicted marginal means with 95% confidence intervals (CI) adjusted for these potential confounders were reported for linear regressions. Adjusted, predicted marginal percentages and odds ratios (OR) with 95% CI were reported for logistic regression models. Non-overlapping 95% CIs between smoking groups were interpreted as statistically significant.

Results

Descriptive Statistics

Of the 4,848 adolescents with complete data who participated in the MEC component of the NHANES between 2001 and 2008, about 25% were sampled in each of the four 2-year cycles, 50% were male, 50% 12–14 years old, 63% white non-Hispanic, 14% black non-Hispanic, and 18% Hispanic (Table 1). Adolescents were oversampled in NHANES from 2001–2006, but this oversampling was discontinued for the 2007–2008 cycle, accounting for the smaller number of adolescents in this period. Never smokers constituted 65.7% (n = 3,158) of the participants, not current smokers, 20.5% (n = 1,085), and current smokers, 13.8% (n = 605). The percentage of current smokers over the study period did not change: 13.9% (95% CI 10.9–17.7) in 2001–2002, 13.4% (95% CI 10.6–16.8) in 2003–2004, 14.5% (95% CI 12.1–17.2) in 2005–2006, and 13.4% (95% CI 11.4–15.5) in 2007–2008 (Figure 1). However, the percentage of never smokers increased significantly from the 2001–2002 cycle (60.2%, [95% CI 55.8–64.4]) to the 2007–2008 cycle (70.5%, [95% CI 68.0–73.0]).

Adolescents reported, on average, 2.0 (95% CI 1.8–2.2) physically unhealthy days, 2.6 (95% CI 2.4–2.9) mentally unhealthy days, and .95 (95% CI 0.79–1.11) activity limitation days. Approximately 8% of the sample reported fair or poor health.

Linear Regression Analyses

For the baseline model of physically unhealthy days, NHANES cycle, gender, age, race-ethnicity, and PIR were not significantly associated with physically unhealthy days. After controlling for these variables in the two-smoking group analyses, ever smokers reported more physically unhealthy days (2.5 days [95% CI 2.1–2.9]) than did never smokers (1.8 days [95% CI 1.5–2.1]; p < .001) (Table 2). Although the three-smoking group comparison was statistically significant overall (p < .01), all the 95% CI for the adjusted mean physically unhealthy days among the three groups overlapped, suggesting nonsignificant differences among these groups.

For the baseline model of mentally unhealthy days, NHANES cycle, gender, age, race-ethnicity, and PIR were all statistically significantly associated with mentally unhealthy days. Adolescent females reported 1.33 more mentally unhealthy days than adolescent males; adolescents aged 15–17 years reported 1.06 more mentally unhealthy days than younger adolescents aged 12–14 years; black non-Hispanic and Mexican American Hispanic adolescents reported fewer mentally unhealthy days than white non-Hispanic adolescents; adolescents from poorer families reported more mentally unhealthy days than those from wealthier families; and individuals assessed during 2007–2008 data collection cycle reported more mentally unhealthy days than those during other NHANES data cycles. After adjustment for these potential confounders, in the two-group smoking group analyses, individuals who ever smoked reported 75% more mentally unhealthy days (3.66 days [95% CI 3.12–4.20]) than those who never smoked (2.08 days [95% CI 1.81–2.35]; p < .0001). Similarly, in the three-smoking group analyses, current smokers (4.21 days [95% CI 3.50–4.91]) and not current smokers (3.32 days [95% CI 2.60–4.03]) reported more mentally unhealthy days than never smokers (2.08 days [95% CI 1.81–2.35]; overall p < .0001).

For the baseline model for activity limitation days, only the NHANES data cycle was statistically significant, with adolescents assessed during the 2007–2008 data cycle reporting more activity limitation days than those assessed in earlier cycles (p < .001). After adjustment for potential confounders in the two-smoking group analyses, ever smokers reported more activity limitation days (1.29 days [95% CI .98–1.59]) than did never smokers (0.78 days [95% CI .62–.94]; p < .01). For the three-smoking group analyses, current smokers reported significantly more activity limitation days (1.57 days [95% CI 1.11–2.02]) than never smokers (.77 days [95% CI .61–.93]; overall p < .01) but not significantly more than not current smokers (1.11 days [95% CI .79–1.43]).

Logistic Regression Analysis

None of the potential confounders was significantly associated in a baseline model with frequent physical distress, similar to the results for the baseline model for physically unhealthy days. For the adjusted two-smoking group model, those who ever smoked were more likely to report frequent physical distress than those who never smoked (OR = 1.75 [95% CI 1.22–2.52]). For the three-smoking group model, both those who were not current smokers (OR = 1.62 [95% CI 1.01–2.59]) and those who were current smokers (OR = 1.99 [95% CI 1.20–3.32]) were more likely to report frequent physical distress than those who never smoked.

The same potential confounders significantly associated in the baseline model for mentally unhealthy days (NHANES cycle, gender, age, race-ethnicity, and PIR) were also statistically significantly associated with frequent mental distress. For the adjusted two-smoking group model, those who ever smoked were more likely to report frequent mental distress than those who never smoked (OR = 2.08 [95% CI 1.44–3.01]). For the three-smoking group model, both those who were not current smokers (OR = 1.63 [95% CI 1.03–2.59]) and those who were current smokers (OR = 2.84 [95% CI 1.80–4.49]) were more likely to report frequent mental distress compared with those who never smoked, respectively.

For the multivariate baseline model, Mexican-American adolescents reported fewer activity limitations than adolescents in other racial-ethnic groups. For the adjusted two-smoking group model, those who ever smoked did not differ statistically significantly in frequent activity limitations from those who ever smoked. For the three-smoking group model, only current smokers were more likely to report frequent activity limitations than those who never smoked. (OR = 2.65 [95% CI 1.38–5.08]).

Finally, in the baseline model, gender, PIR, and NHANES cycle were all statistically significantly associated with fair or poor self-rated health. Adolescent girls were more likely to report fair or poor health than adolescent boys; adolescents from lower income households were more likely to report fair or poor health than adolescents from higher income households; Adolescents assessed during the 2007–2008 NHANES data collection cycle more likely reported fair or poor health than those in earlier cycles (p < .001). For the adjusted two-smoking group model, those who ever smoked were statistically significantly more likely to report fair or poor health than those who never smoked (OR = 1.42 [95% CI 1.05–1.92]). Although the three-smoking group model showed no overall statistically significant differences among the groups, those who currently smoked were still statistically significantly more likely to report fair or poor health relative than those who never smoked (OR = 1.61 [95% CI 1.09–2.37]).

When we tested for effect modification for both the two-group and three-group smoking status variables by gender and race and ethnicity, we found no statistically significant effects. In the analysis where we tested the associations between serum cotinine and HRQOL, we found very similar associations to those found with the smoking status exposure variables. The continuous log-transformed serum cotinine estimates were associated with both physically (p < .03) and mentally unhealthy days (p < .0001). Examination of HRQOL with serum cotinine level dichotomized into ≤10.0ng/mL versus >10.0ng/mL did result in statistically significant results in the expected direction but were smaller due to the decrease in the power of the tests. The result for physically unhealthy days was not statistically significant, while the result for mentally unhealthy days was significant at p < .01. When we adjusted for the smoking status variables, there were no statistically significant associations between serum cotinine and either of the HRQOL outcomes, further validating that the two types of smoking status variables measure similar constructs.

Discussion

This is the first study that we are aware of that assessed the associations between youth smoking, defined using both self-reports and serum cotinine levels, and four measures of HRQOL. Although smoking is clearly associated with worse HRQOL among adults (McClave et al., 2009), the extent to which smoking is associated with worse HRQOL among adolescents is unclear. This study provides new evidence regarding the link between cigarette smoking and HRQOL among a representative sample of adolescents aged 12–17 years. Our findings are consistent with previous studies among adults and the few studies conducted among adolescents demonstrating that youth who smoke cigarettes are more likely to report poor HRQOL compared with youth who never smoked. The significant associations between smoking and HRQOL further support the hypothesis that smoking has potentially significant short-term and long-term health consequences for many different health domains among adolescents.

Adolescence in itself is a vulnerable period of human development when experimentation and autonomy emerge (Schulenberg, Maggs, & Hurrelman, 1999). Experimentation and the use of tobacco products increase the risk for establishing a more regular pattern of use as well as the likelihood for health-related problems. In this study, adolescents who ever tried smoking or reported current smoking had poor opinions of both their physical and mental health. Adolescents who had ever tried smoking were significantly more likely to report physically and mentally unhealthy days, activity limitation days, and fair or poor general health than never smokers. General self-rated health provides a global assessment of recent and past health that includes perceptions of health beyond just physical and mental health. Importantly, the relationship between youth smoking and activity limitations has not been shown previously in the literature for adolescents. Activity limitation represents a class of more severe morbidity outcomes and for adolescents, likely related to academic performance such as attention and learning, and mental health outcomes related to anxiety and depression. Smoking contributes to approximately $193 billion in healthcare and lost productivity each year and 5.1 million years of potential life lost in the United States annually (Centers for Disease Control and Prevention, 2008). Thus our findings suggest that the association between smoking and HRQOL may begin to take a toll early in the lifespan.

For the three-smoking group comparisons, a dose-response relationship between smoking status and mentally unhealthy days was observed and is similar to that in a previous study among adult students aged 18–24 years (Zahran, Zack, Vernon-Smiley, & Hertz, 2007). In this study, adolescents who were current smokers reported two or more mentally unhealthy days than those who never smoked. Given that a one-day difference in HRQOL days is considered meaningful (Centers for Disease Control and Prevention, 2000; Moriarty et al., 2005), our findings suggest that adolescents may be attempting to ameliorate negative emotional experiences through cigarette use. Previous evidence also suggests that early childhood trauma affects psychosocial development, which in turn may cause individuals to be more likely to experiment with substances such as cigarettes as a means to ameliorate these negative feelings (Anda et al., 1999; Dube et al., 2003). While trauma-related histories are not available in this sample, our findings are consistent with other studies that have documented relationships between adverse childhood experiences, poor mental health, stress, depressed mood, and smoking among adolescents (Anda et al., 1999; Degenhardt & Hall, 2001; DiFranza et al., 2004; Dube et al., 2010; Jasuja et al., 2008; Jorm et al., 1999; Jun et al., 2008; van Loon et al., 2005; Lawrence et al., 2010; Tercyak et al., 2002).

The assessment of HRQOL in relationship to smoking among adolescents may add to the growing body of literature that attempts to understand the relationships and processes related to youth risks and assets and their decisions to engage or not engage in high-risk behaviors, such as tobacco use (Brook, Rubenstone, Zhang, Morojele, & Brook, 2011; Fulkerson et al., 2006; Murphey, Lamonda, Carney, & Duncan, 2004; Oman et al., 2004; Scales, Leffert, & Vraa, 2003).This promising field may give rise to interventions and health promotion programs that decrease the vulnerability of youth to initiate smoking and that improve poor HRQOL. For example, rather than engaging in risk behaviors such as tobacco use, youth can become more actively engaged in preventing tobacco use. Centers for Disease Control and Prevention’s Best Practices User Guide: Youth Engagement—State and Community Interventions (2010d) focuses on the role youth play in advancing policy as part of a comprehensive tobacco control program. This type of youth engagement has the potential to decrease smoking behaviors (Winkleby, Feighery, Dunn, Kole, Ahn, & Killen, 2004) and improve overall quality of life as a result. Interventions such as youth engagement should be a coordinated effort with comprehensive tobacco control strategies known to prevent and reduce tobacco use (Centers for Disease Control and Prevention, 2007).

Continued efforts to implement comprehensive tobacco control programs and policies effective in preventing and curbing tobacco use among youth are needed. However, understanding the psychosocial dynamics of smoking among this population is especially critical given that tobacco product marketing has been targeted toward vulnerable populations, such as youth (John, Cheney, & Azad, 2009). Moreover, adolescents with poor self-rated health may be a group especially vulnerable to tobacco industry marketing. In fact, adolescents with depressive symptoms and high receptivity to tobacco advertisements are likely to engage in smoking (Tercyak et al., 2002). Future studies on youth tobacco use would benefit from the inclusion of measures on HRQOL to assess whether youth with poor HRQOL are more receptive to pro-tobacco influences.

Other important points of intervention for addressing smoking among adolescents include pediatricians, school health professionals, and mental health providers. The American Academy of Pediatrics (AAP) has indicated that tobacco is a substance of abuse (Sims & Abuse, 2009), and it is critical for healthcare providers to screen for tobacco use and to emphasize and warn about the dangers of initiating any type of tobacco use, as well as to identify youth who may need cessation treatments. This study reinforces the importance of including an assessment of tobacco use and smoking status among adolescents in the healthcare setting, especially as the behavior may lead to further experimentation and use and may also be related to concomitant physical and particularly mental dysfunction.

Our study is subject to several limitations. First, temporal and causal relationships between smoking and HRQOL among youth cannot be made because of the cross-sectional design of NHANES. Second, the group who reported ever smoking one or more cigarettes but not currently does not distinguish established smokers who quit from experimenters, which is important to consider as adolescence is a period when smoking initiation occurs. Third, even though study participants reported their smoking behaviors privately using a computer, they may not have reliably reported their smoking behavior; however, the NHANES cotinine data allowed us to increase the reliability of classifying current smokers. In fact, in our study, the prevalence of current smoking among youth was higher at 13.4% in 2007–2008 NHANES than that in the self-reported National Survey on Drug Use and Health (10.5% for 2007 and 9.8% for 2008; Substance Abuse and Mental Health Administration, 2009). Fourth, because the physical healthy days question includes a reference to injuries, the reported results may be biased toward the null and underestimate the true association. Fifth, it is possible that other adolescent risk behaviors that were not included in our analysis could have affected the strength of the estimates (either upward or downward) of the associations between smoking and HRQOL. Finally, because the HRQOL measures are broad self-reported measures of health, they may not be sensitive enough to detect associations with more specific health outcomes that could be assessed and measured with an in-person clinical evaluation.

Despite these limitations, our study demonstrated a clear association between smoking and poor HRQOL, including activity limitations, among youth. Taking an upstream approach requires not only implementing and sustaining comprehensive tobacco control policies and programs known to prevent youth tobacco use but also requires understanding and addressing adverse mental and physical well-being that may be associated with youth tobacco use. Finally, our findings on the relationship between poor HRQOL and smoking among adolescents underscores the importance of addressing concomitant health issues among adolescents.

Declaration of Interests

All authors of this submission, Shanta R. Dube, William Thompson, David Homa, and Matthew Zack have no conflicts of interest.

Funding

None declared.

Acknowledgments

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

References