- Email: [email protected]
Anxiety and Depressive Disorders Are Associated with Smoking in Adolescents with Asthma
Journal of Adolescent Health 40 (2007) 425– 432
Original article
Anxiety and Depressive Disorders Are Associated with Smoking in Adolescents with Asthma Terry Bush, Ph.D.a,*, Laura Richardson, M.D., M.P.H.b, Wayne Katon, M.D.c, Joan Russo, Ph.D.c, Paula Lozano, M.D., M.P.H.b, Elizabeth McCauley, Ph.D.c, and Malia Olivera a
Center for Health Studies, Group Health Cooperative, Seattle, Washington Department of Pediatrics, University of Washington Medical School, Seattle, Washington c Department of Psychiatry & Behavioral Sciences, University of Washington, Seattle, Washington Manuscript received September 18, 2006; manuscript accepted November 30, 2006 b
Abstract
Purpose: To evaluate the association between mental health indicators (including meeting criteria for one or more DSM-IV [Diagnostic and Statistical Manual of Mental Disorders—fourth edition] anxiety or depressive disorders) and susceptibility to smoking or current smoking among youth with asthma and to evaluate the impact of smoking on asthma symptoms and self-management. Methods: We conducted telephone interviews with a population-based sample of 11- to 17-yearold youth and their parents (n ⫽ 769). Interview content included questions on smoking behaviors, asthma symptoms and treatment, externalizing behavior, and a structured psychiatric interview to assess DSM-IV anxiety and depressive disorders. Results: Five percent of youth were smokers and 10.6% indicated that they were “susceptible to smoking.” Smoking was more common among youth with mental health disorders. Anxiety/ depressive disorders were present in 14.5% of nonsmokers, 19.8% of susceptible nonsmokers, and 37.8% of smokers. After controlling for important covariates, youth with more than one anxiety and depressive disorder were at over twofold increased risk for being a smoker. Similarly, for each one-point increase in externalizing disorder symptoms, youth had a 10% increase in likelihood of being a smoker and a 4% increase in risk for “susceptibility to smoking.” Youth who were smokers reported more asthma symptoms, reduced functioning due to asthma, less use of controller medicines, and more use of rescue medications. Conclusions: Comorbid mental health disorders are associated with increased risk of smoking in youth with asthma. Smoking is associated with increased asthma symptom burden and decreased controller medication use. Interventions for youth with asthma should consider screening for and targeting these behavioral concerns. © 2007 Society for Adolescent Medicine. All rights reserved.
Keywords:
Adolescent; Asthma; Smoking; Mental disorder classification; Susceptibility; Anxiety/depression
With a prevalence of up to 10%, asthma is the most common medical disorder among adolescents [1]. When present, asthma is associated with increased functional impairment and lost days at school and increases the risk for development of anxiety or depressive disorders [2,3]. *Address correspondence to: Dr. Terry Bush, Center for Health Studies, Group Health Cooperative, 1730 Minor Ave, Suite 1600, Seattle, WA 98101-1448. E-mail address: [email protected]
Smoking is also common in adolescents, with prevalence rates estimated at 8.3% among middle school students and 21.8% among high school students [4]. Among youth with asthma, prevalence estimates of smoking have been reported to be similar to or higher than those of the general population [5–7]. Smoking and exposure to secondhand smoke are associated with an increased requirement for asthma medication [8], more emergency room visits, and more hospital admissions [9]. Despite known risks, many youth with asthma continue to experiment with smoking, but factors associated with
1054-139X/07/$ – see front matter © 2007 Society for Adolescent Medicine. All rights reserved. doi:10.1016/j.jadohealth.2006.11.145
426
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
smoking onset are not well understood. Common risk factors for youth smoking in general include anxiety [10] and depressive [11] disorders and parental smoking [5,6,12]. Some studies suggest that adolescents with asthma have reasons for smoking similar to those of the general public [7,13]. Common risk factors for smoking in adolescents with or without asthma include exposure to smoking role models, perceived parental approval of smoking, perceived social value of smoking, and rebelliousness [6]. However, a recent study did not find an association between depressive symptom measures and smoking progression over time in youth with asthma [13]. One limitation of this study is that it did not have information on common mental health comorbidities (such as anxiety and externalizing disorders) and lacked information on Diagnostic and Statistical Manual of Mental Disorders—Fourth edition (DSM-IV) diagnoses. Additionally, prior studies did not have data on asthma severity, symptom burden or treatment, which may be important confounders of the relationship between mental health disorders and smoking in addition to understanding the potential impact of smoking on quality of life for these youth. Another limitation of prior studies among youth with asthma is that they do not assess susceptibility to smoke. This is an important variable to measure because the onset of smoking begins in youth with a lack of a firm resolve to not smoke and progresses through stages of susceptibility to smoking including experimentation and occasional use to regular smoking [12,14]. Although smoking in young teens is rare, susceptibility to smoke is more common and is a strong predictor of future smoking [15]. The goals of this study are to evaluate, among youth with asthma, the association between smoking behaviors and mental health symptoms and disorders (externalolizing, anxiety and depressive) and also to evaluate the association between smoking and asthma symptoms and selfmanagement.
Methods We used data from a large epidemiological study of the prevalence of mental health disorders among adolescents with asthma [16] (unpublished data, 2005). Participants were initially identified from automated pharmacy and registration data of Group Health Cooperative (GHC), a large nonprofit health maintenance organization with over 600,000 enrollees in urban and rural settings in Washington State representing about 10% of the state population. GHC enrollees are largely representative of western Washington with the exception of fewer with higher levels of income. Youth were eligible for the study if they had been enrolled in the health plan for at least 6 months, were between 11 and 17 years of age, and had asthma defined as having at least
one of the following in the past 12 months based on review of automated data: hospitalized with an asthma diagnosis and at least one dispensing of an asthma medication during the same time period; more than one emergency room or urgent care visit with an asthma diagnosis; at least two office visits with an asthma diagnosis and at least one dispensing of an asthma medication; only one visit with an asthma diagnosis but at least two dispensings of asthma medications filled on different days; at least four dispensings of asthma medications; at least one office visit with an asthma diagnosis and another in the past 18 months and at least one dispensing of an asthma medication. Youth with evidence of bipolar disorder or schizophrenia based on International Classification of Diseases—ninth version (ICD-9) codes from automated data were excluded. Parents and eligible youth were contacted by phone and invited to participate in the study. Parents were eligible for the adult survey if they were the parent or legal guardian who had the most knowledge of their child’s health and health care. Interviewers from the Center for Health Studies Survey Department obtained verbal informed consent and conducted separate telephone interviews with parents and youth. Interviewers received over 12 hours of training by survey professionals and the study clinician (E.M.). Interview quality and reliability was assured through ongoing monitoring, analysis, feedback, and corrective instructions if needed. Interviewers were blinded to the hypotheses of this study. Participating youth received a $25 coupon to a local store after completion of the youth interview. The Institutional Review Board of Group Health Cooperative approved all study procedures. Among the 1288 families eligible for the study, 769 parents (59.7%) and 781 youth (60.6%) completed the interview (Figure 1). After excluding parents who refused access to youth medical records, our weighted sample size was 757 parents and 769 youth. Comparisons between survey responders and nonresponders revealed that nonresponders were more likely to live in a large town (p ⬍ .02) but less likely to be on Medicaid (p ⬍ .02), to have made a mental health visit (p ⬍ .01), to have received a depression diagnosis (p ⬍ .01), or to have filled a prescription for antidepressants in the past year (p ⬍ .01). Measures Mental health measures The main variables examined in these analyses were presence of DSM-IV anxiety and depressive disorders based on interviews conducted with youth using the Diagnostic Interview Schedule for Children NIMH DISC-4.0 (C-DISC) [17] and externalizing behaviors assessed during the parent
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
427
Figure 1. Recruitment for the Youth with Asthma Study
interview using the Child Behavior Checklist (CBCL) [18]. The C-DISC has been shown to be a reliable and valid structured interview to diagnose DSM-IV disorders in children and adolescents [17]. We administered the C-DISC with adolescents and the CBCL with parents because research indicates that youth are more accurate at reporting symptoms of an internalizing disorder and parents are more accurate at reporting youth externalizing behaviors [19]. We administered the C-DISC depression modules (major depression, dysthymia) and anxiety modules (panic disorder, separation anxiety, social phobia, generalized anxiety and agoraphobia). Analyses were initially conducted separately for major depression, panic disorder, and anxiety disorders. Because cross-sectional and longitudinal data consistently show a high correlation between anxiety and depression, we also created and evaluated a separate variable describing youth with at least one anxiety or depression diagnosis versus none. Smoking measures We assessed adolescent smoking with three standard and validated questions commonly used in youth surveys: 1) “have you ever tried or experimented with cigarette smoking, even a few puffs?,” 2) “have you smoked at least 100 cigarettes in your life?,” and 3) “think about the last 30 days. On how many of these days did you smoke?” [13,20]. The primary outcome measures for this article were current smoking (defined as having smoked at least 100 cigarettes and having smoked in the past month) [20] and susceptibility to smoking [14]. We assessed susceptibility to smoking for nonsmokers with three standard questions: 1) “Do you think you will try a cigarette soon?,” 2) “If one of your best friends were to offer you a cigarette, would you smoke it?,” and 3) “Do you think you will be smoking one year from now?” The response option for questions 2 and 3 is “defi-
nitely yes,” “probably yes,” “probably not,” or “definitely not.” As defined, a person was considered susceptible to smoke if they thought they would try a cigarette soon or if they answered anything other than ‘definitely not’ to questions 2 and 3 [14]. We calculated the total number of youth who were either susceptible to smoking or who were current smokers, because both increase the risk for future smoking. We also asked youth about household smoking with the following, “Has anyone in the family smoked in the last year, including smoking outside of the house?” Asthma measures The diagnosis of asthma for youth was confirmed with parents during their interview. Youth were asked about the age of onset of their asthma and the number of symptom days in the prior 2 weeks. Youth also completed three scales from the Children’s Health Status-Asthma Teen version (CHSA-T), measuring asthma-specific emotional and physical functioning and activity limitations [21]. We used automated diagnostic, visit, and pharmacy data to measure asthma severity based on the Health Employer Data and Information Set (HEDIS) definition of children/ adolescents who are at high risk of adverse asthmatic events [22]. In this definition, meeting criteria for any of the following four health care utilization variables over a 12month period identified individuals at high medical risk: 1) four or more dispensings of asthma medication; 2) one or more emergency visits with an asthma diagnosis; 3) one or more hospitalizations with an asthma diagnosis; 4) four or more ambulatory visits with an asthma diagnosis. Approximately 66% of our sample met the criteria for four or more dispensings of an asthma medication, suggesting lack of specificity as a measure of severity in our population. We therefore developed a second HEDIS measure with a modified medication criteria (replacing criteria 1 above with the
428
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
requirement of having one or more oral steroid prescriptions for asthma in the past 12 months). We tested both of these measures in separate analyses and found similar results. Table 1 shows results using the original HEDIS measure. We also compared smoking groups on use of rescue versus controller medications. Rescue medications such as albuterol are used as needed for patients who continue to have breakthrough symptoms despite use of controller medications. Controller medicines such as fluticasone propionate are intended to be taken on a regular basis to prevent symptoms. The distinction is important because smoking is associated with resistance to medicines, which may lead to increased use of rescue medicines for relief of symptoms. Demographics Demographic variables were evaluated as covariates in the main analysis. Parents were asked about their race, ethnicity, marital status, and the education and employment status of both parents. Children were asked to confirm their age and gender. To assess socioeconomic status, we used the GEOCODE, a measure of neighborhood median household income based on subject zip code. Rural Urban Commuting Area (RUCA) codes were used to further categorize household location as rural or urban. We also included a variable comparing youth from eastern versus western Washington because of the distinct differences between these two areas of the state on population density, environmental exposures to tobacco, socioeconomic status, and health care access. Because of these differences, we hypothesized that there may also be geographic differences in youth smoking rates. Chronic disease measure Chronic diseases other than asthma may be associated with smoking and mental health concerns. To evaluate for possible confounding, we used a modified form of the Pediatric Chronic Disease Scale (PCDS). The PCDS is a measure of medical comorbidity developed based on automated pharmacy data at Group Health and has been shown to correlate with subsequent health care costs and hospitalization rates [23]. For this study, the PCDS was modified by removing medications that are commonly used for asthma and anxiety or depression.
of Diseases—ninth version (ICD-9) diagnosis of depression or anxiety. In conducting this propensity analysis, we first used logistic regression to estimate a response probability for each survey respondent (response propensity score). Then, we calculated weights that were inversely proportional to estimated probability of response rescaled to sum to the observed sample size (i.e., the number of survey respondents). In weighted analyses, individuals with a low probability of response were given higher weights to represent the larger number of nonrespondents with similar characteristics. For this report, all regression analyses were conducted with propensity score weighting. First, we conducted univariate analyses comparing all demographic, asthma, and mental health variables among the three smoking groups. For these analyses, we used chi-square tests with corrections for continuity and analyses of variance (ANOVAs) for categorical and continuous variables, respectively. In the event of a significant global group difference, we conducted two planned tests (either ANOVA or chi-square) to determine which set of group differences was responsible for the statistical significance: nonsmokers versus smokers or nonsmokers versus susceptible nonsmokers. To determine which factors were independently and significantly associated with susceptibility to smoking and current smoking in comparison to nonsmoking, we conducted propensity weighted logistic regression analyses. Two sets of regression analyses were run. One compared the smoking with the nonsmoking group and a second compared the susceptible to smoking with the nonsmoking group. All factors that were statistically significant in bivariate analyses (p ⬍ .05) were initially entered into the models. Then, individually, the least significant variable (the variable with the largest p-value that was greater than .05) was eliminated and the models were refit. This was continued until the models contained only statistically significant (p ⬍ .05) factors. We then estimated the odds ratios (OR) and 95% confidence intervals (CI) for these variables. To make sure that no other significant variables were omitted from the models, the significance of all variables added to the model were assessed.
Analysis
Results
Using de-identified data, we conducted a nonresponder analysis to determine if survey respondents differed significantly from nonrespondents. We used propensity analysis to estimate the probability of being a respondent as a function of age, gender, RUCA code (rural versus urban area), being on Medicaid or state-funded insurance because of low income, PCDS, number of primary care visits, number of asthma-related emergency room visits and hospitalizations, oral steroid prescription, number of specialty mental health visits, any prescription for antidepressant medications or anti-anxiety medications, and an International Classification
Table 1 presents study variables for the total sample completing the interview as well as study variables stratified by youth smoking status. Among participating parents, 46% were female, 92.1% were educated beyond high school, 93.8% were employed, 72.9% were married, and 80.2% were white. Among participating adolescents (mean age ⫽ 14 years), 46.6% were female, 47.3% reported exposure to household smokers, 71.2% were on an asthma controller medication, and 16.2% met DSM-IV criteria for one or more anxiety or depressive disorders. As shown in Table 1, 15.6% of youth were either current smokers or susceptible
Table 1 Characteristics of adolescents with asthma by smoking status Current smoker n ⫽ 38 (5.0%)
Nonsmoker susceptible n ⫽ 82 (10.6%)
Nonsmoker not susceptible n ⫽ 649 (84.4%)
Overall 2 or F test
Current vs. Nonsmoker
Susceptible vs. Nonsmoker
60 (7.9%) 710 (93.8%) 551 (72.9%) 606 (80.2%) 104 (13.5%)
7 (19.4%) 31 (86.1%) 15 (41.7%) 27 (73.0%) 3 (7.9%)
9 (11.1%) 77 (95.1%) 62 (76.5%) 62 (76.5%) 12 (14.8%)
44 (6.9%) 602 (94.1%) 474 (74.2%) 517 (81.0%) 89 (13.7%)
p ⬍ .01 ns p ⬍ .001 ns ns
p ⬍ .05 ND p ⬍ .001 ND ND
ns ND ns ND ND
50,460 ⫾ 17,015
49,385 ⫾ 13,884
51,962 ⫾ 20,418
50,322 ⫾ 16,696
ns
ND
ND
359 (46.6%) 14.0 ⫾ 1.9
28 (73.7%) 16.1 ⫾ .8
33 (40.2%) 14.4 ⫾ 1.6
298 (45.8%) 13.8 ⫾ 1.9
p ⬍ .01 p ⬍ .001
p ⬍ .001 p ⬍ .001
ns p ⬍ .05
213 (27.7%) 234 (30.4%) 323 (41.9%) 362 (47.3%)
0 (0.0%) 1 (2.6%) 37 (97.4%) 33 (86.8%)
11 (13.4%) 34 (41.5%) 37 (45.1%) 51 (63.0%)
202 (31.1%) 199 (30.6%) 249 (38.3%) 285 (43.0%)
p ⬍ .001
p ⬍ .001
p ⬍ .01
p ⬍ .001
p ⬍ .001
p ⬍ .001
ns
ND
6.7 ⫾ 4.3
7.8 ⫾ 5.1
7.3 ⫾ 4.4
6.6 ⫾ 4.2
3.8 ⫾ 4.0 83.7 ⫾ 17.1 83.7 ⫾ 14.6 73.7 ⫾ 15.6 .40 ⫾ .70
6.3 ⫾ 5.2 76.8 ⫾ 19.2 79.0 ⫾ 16.1 65.4 ⫾ 14.7 .64 ⫾ .96
4.4 ⫾ 4.6 83.6 ⫾ 18.7 83.3 ⫾ 14.4 71.9 ⫾ 16.2 .38 ⫾ .63
3.6 ⫾ 3.8 84.1 ⫾ 16.7 84.2 ⫾ 14.5 74.4 ⫾ 15.4 .39 ⫾ .69
p ⬍ .001 p ⬍ .05 ns p ⬍ .001 ns
p ⬍ .001 p ⬍ .01 ND p ⬍ .01 ND
ND ns p ⬍ .07 ND ns ND
614 ⫾ 960
859 ⫾ 1294
614 ⫾ 806
600 ⫾ 954
ns
ND
ND
p ⬍ .05 (df ⫽ 4)
p ⬍ .05
ns
26 (3.4%) 197 (25.6%) 548 (71.2%)
0 (0.0%) 16 (41.0%) 23 (59.0%)
4 (4.9%) 26 (31.7%) 52 (63.4%)
22 (3.4%) 155 (23.8%) 473 (72.9%)
9.2 ⫾ 8.2 56 (8.0%) 19 (2.5%) 104 (13.6%)
17.4 ⫾ 11.8 6 (20.7%) 2 (5.6%) 11 (29.7%)
11.5 ⫾ 9.3 10 (13.3%) 3 (3.7%) 14 (17.3%)
8.5 ⫾ 7.6 40 (6.7%) 14 (2.2%) 79 (12.2%)
p ⬍ .001 p ⬍ .01 ns p ⬍ .01
p ⬍ .001 p ⬍ .01 ND p ⬍ .01
p ⬍ .001 ns ND p ⬍ .001
124 (16.2%)
14 (37.8%)
16 (19.8%)
94 (14.5%)
p ⬍ .001
p ⬍ .001
ns
429
ND ⫽ not determined because the overall test was not significant; HEDIS ⫽ Health plan Employer Data and Information Set; PCDS ⫽ pediatric chronic disease score; MH ⫽ mental health diagnosis. a Children’s Health Status-Asthma Teen version (CHSA-T). b Modified HEDIS revealed similar results. c Parent rated (Child Behavior Checklist).
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
Parent demographics Highest education level high school or less, either parent Employed full or part time, either parent % Married % Caucasian, participating parent % Medicaid GEOCODE (mean household income based on zip code) Youth demographics % female Mean age ⫾ SD Age groups (years) % 11–12 % 13–14 % 15–17 % Households with smoker (past year) Asthma variables Age onset with asthma Asthma severity: Symptom days past 2 weeks Health status–Activitiesa Health status–Emotional Health status–Physical HEDIS Measure (0–4)b Medical comorbidity score (PCDS w/o asthma, w/o a MH diagnosis) Asthma treatment intensity No medications Albuterol only At least 1 controller Mental health Externalizing scalec % With major depression % With panic disorder % With any anxiety disorder % With 1⫹ anxiety or depression disorders
Total n ⫽ 769 (100%)
430
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
to smoking in the future. Although current smoking was present in only 5%, it was more common among older adolescents (11.5% among 15–17-year-olds and less than 1% [n ⫽ 1] in 11–14-year-olds). Smokers significantly differed from nonsmokers on many important factors. In comparison with nonsmokers, smokers were significantly older, more likely to be female, less likely to come from homes where parents were married and educated beyond high school, and more likely to report exposure to household smoking. Smokers also reported more asthma symptom days in the past 2 weeks and more functional limitations (CHSA-T activities and physical health scales). Smokers were more likely to use rescue medications such as albuterol and less likely to use controller medications despite having more asthma symptom days per week. Each of the mental health disorder categories examined were more common among smokers: smokers were more likely to meet DSM-IV diagnostic criteria for a depressive or anxiety disorder, and had higher parent-rated externalizing disorder symptoms. Compared with nonsusceptible, nonsmoking youth, susceptible nonsmoking youth were older, more likely to have a family member that smoked, had higher mean externalizing symptom scores, and were more likely to have an anxiety disorder. The multivariate logistic model controlling for significant covariates (Table 2) showed that, in comparison to nonsusceptible nonsmoking youth, those who smoked were significantly more likely to be older (OR 2.90, 95% CI 1.98 – 4.26) and female (OR 2.40, 95% CI 1.01–5.72) and to have one or more family members who smoked (OR 6.59, 95% CI 2.37–18.31). They had significantly higher parentrated externalizing behavior symptom scores (OR 1.10, 95% CI 1.05–1.14) and were more likely to meet DSM-IV criteria for one or more anxiety or depressive disorders (OR 2.58, 95% CI 1.06 – 6.28). Compared with nonsusceptible nonsmoking youth, youth who were susceptible to smoking (Table 3) were significantly older (OR 1.18, 95% CI 1.04 –1.34), more likely to have one or more family members who smoked (OR 1.92, 95% CI 1.17–3.14), and had higher externalizing scale scores (OR 1.04, 95% CI 1.01–1.07). After controlling for
Table 2 Weighted logistic regression for adolescent smokers versus the nonsmokers Characteristic
Odds ratio
95% CI
p
Youth age Household smoking Female (youth) CBCL externalizing behavior DSM-IV anxiety-depressive disorder
2.904 6.591 2.399 1.097
(1.978–4.262) (2.372–18.314) (1.006–5.721) (1.053–1.144)
.000 .000 .048 .000
2.577
(1.057–6.284)
.037
CI ⫽ confidence interval; CBCL ⫽ Child Behavior Checklist; DSM-IV ⫽ Diagnostic and Statistical Manual– 4th edition.
Table 3 Weighted logistic regression for the susceptible versus nonsusceptible nonsmokers Characteristic
Odds ratio
95% CI
p
Youth age Household smoking CBCL externalizing behavior
1.179 1.920 1.037
(1.040–1.337) (1.172–3.143) (1.010–1.065)
.010 .010 .008
CI ⫽ confidence interval; CBCL ⫽ Child Behavior Checklist.
covariates, anxiety disorders were no longer significantly associated with susceptibility to smoke.
Discussion In this population-based sample of youth with asthma, we found that youth with comorbid anxiety and depressive disorders or elevated externalizing symptoms were more likely to be smokers. These associations persisted even after controlling for potentially important demographic and clinical covariates. Adolescents with asthma who smoked were more than twice as likely to have major depression (20.7% vs. 6.7%) and one or more anxiety disorders (29.7% vs. 12.2%) compared with nonsmokers. This finding is consistent with studies of youth respondents from the community where anxiety and depressive disorders have been shown to be associated with increased risk for smoking [10,11]. Our findings are also consistent with increasing evidence in cross-sectional and longitudinal studies of a link between panic attacks, depression and anxiety disorders, and smoking [10,11,24 –26]. The results of this study are in contrast with a recent study of youth with asthma which found that depressive symptoms or sadness did not increase the risk for continued smoking [13]. One reason our study results may differ is that we looked at the association cross-sectionally rather than longitudinally. We also had more specific criteria for delineating mental health disorders, including the use of structured psychiatric interviews and parental ratings of externalizing symptoms. Because of the cross-sectional nature of our data, we cannot determine the direction of association between smoking and anxiety or depressive disorders. Prior studies suggest that the relationship between smoking and mental disorders may be bidirectional [10]. Analysis of longitudinal data indicates that depression or panic [10,27] is associated with a higher risk for onset of smoking and the reverse, that smoking is associated with a higher risk for development of panic [10,27,28] and major depression [10,29,30]. It has also been suggested that the association between mental health factors and smoking may be due to underlying causal factors such as adverse childhood experiences, which are potential antecedents to both depression/ anxiety and smoking [31].
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
To our knowledge, this is the first study to examine externalizing disorder symptoms as a predictor of smoking among asthmatics. However, our finding of an increased risk for smoking in youth with asthma with a high level of externalizing symptoms is consistent with studies in the general population that have shown that youth with externalizing disorders are at increased risk for smoking [32,33]. It is important to note that 87% of the smokers reported at least one family member who smoked. In addition to the risk for the development of smoking, exposure to secondhand smoke in the home increases the risk for development of asthma [34] and may be a contributing factor to the increase in asthma severity we observed in youth smokers. Youth who smoked reported more symptom days, more functional impairment, and more overall asthma symptoms. Although it is unclear if this is a result of patient or health care factors, smokers also reported less use of medications to control asthma and higher use of rescue medications (e.g., albuterol) compared with nonsmokers. Even when youth are using controllers, studies suggest that exposure to tobacco smoke may reduce the effectiveness of asthma medications [8]. One potential limitation of this study is that the prevalence of smoking in our population was low at 5%. This may be due to the young age of our sample. When examining the older age group, the prevalence of smoking was 11.5%. Additionally, it may be due to regional differences in smoking prevalence. The prevalence of smoking is lower in the Pacific Northwest, where this study was conducted, than in other regions of the country. Although smoking was relatively infrequent in our population, twice as many were susceptible to smoking. Studies repeatedly show that susceptibility to smoking and smoking during childhood are strong predictors of future smoking [14,15,35,36]. Given the high risk, these susceptible youth might be a good target for primary prevention. One potential intervention target is media images of smoking. Youth of this age are particularly vulnerable to peer and media influences and begin to formulate positive attitudes about smoking without realizing the addictive properties of nicotine [37]. In a recent longitudinal study, nonsusceptible youth who recalled pro-tobacco messages or role models in print or visual forms (stores, magazines, TV, movies) were at increased risk for subsequent smoking susceptibility [38]. In addition to the low prevalence of smoking and the cross-sectional design, another limitation of this study is the use of a limited number of measures to assess smoking. We recognize there are other ways of measuring and defining smoking and that our measure relies on self-report. Another potential limitation to this study is that the C-DISC was administered only to the youth. It is possible that parentreported anxiety and depression from the DISC may have been different from youth self-report. However, there is controversy in the scientific field about how to include parent data when there is discordance between parent and child data. Research indicates that youth are more accurate
431
at reporting internalizing symptoms than their parents [19]. Strengths of this study include the large population-based sample, the careful assessment of DSM-IV anxiety and depression disorders, and the availability of measures of asthma symptoms and treatment. The results of our study have clinical and public health significance. We have identified important risk factors for smoking and susceptibility to smoke in youth with asthma and provide evidence of the impact of smoking on management of asthma. Clinicians are in a unique position to identify and intervene with youth to encourage quitting or reinforce not smoking [39]. Primary care physicians often screen children/adolescents with asthma for smoking. Our data suggest that it is important to also screen these children/adolescents for anxiety and depressive disorders as well as externalizing behaviors, because of the increased association of these mental health problems with smoking. Efforts to educate the public about the association between smoking and mental disorders are also needed because each has been associated with adverse outcomes among those with asthma [40]. Our results confirm the importance of addressing risk factors for smoking in youth with asthma and developing early prevention and cessation interventions, including efforts to reduce exposure to secondhand smoke. Acknowledgments This work was supported by National Institute of Mental Health grant #MH 67587 (Principal Investigator: Wayne Katon, MD). References [1] Weitzman M, Gortmaker SL, Sobol AM, et al. Recent trends in the prevalence and severity of childhood asthma. JAMA 1992;268: 2673–7. [2] Katon WJ, Richardson L, Lozano P, et al. The relationship of asthma and anxiety disorders. Psychosom Med 2004;66:349 –55. [3] Goodwin RD, Fergusson DM, Horwood LJ. Asthma and depressive and anxiety disorders among young persons in the community. Psychol Med 2004;34:1465–74. [4] Centers for Disease Control and Prevention (CDC). Tobacco use, access, and exposure to tobacco in media among middle and high school students—United States, 2004. MMWR 2005;54:297–301. [5] Otten R, Engels RC, van den Eijnden RJ. Parental smoking and smoking behavior in asthmatic and nonasthmatic adolescents. J Asthma 2005;42:349 –55. [6] Zbikowski SM, Klesges RC, Robinson LA, et al. Risk factors for smoking among adolescents with asthma. J Adolesc Health 2002;30: 279 – 87. [7] Precht DH, Keiding L, Nielsen GA, et al. Smoking among upper secondary pupils with asthma: reasons for their smoking behavior: a population-based study. J Adolesc Health 2006;39:141–3. [8] Strachan DP, Cook DG. Health effects of passive smoking: parental smoking and childhood asthma: longitudinal and case-control studies. Thorax 1998;53:204 –12. [9] Troisi RJ, Speizer FE, Rosner B, et al. Cigarette smoking and incidence of chronic bronchitis and asthma in women. Chest 1995;108: 1557– 61.
432
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
[10] Johnson JG, Cohen P, Pine DS, et al. Association between cigarette smoking and anxiety disorders during adolescence and early adulthood. JAMA 2000;284:2348 –51. [11] Covey LS, Tam D. Depressive mood, the single-parent home, and adolescent cigarette smoking. Am J Public Health 1990;80:1330 –3. [12] Gritz ER, Prokhorov AV, Hudmon KS, et al. Predictors of susceptibility to smoking and ever smoking: a longitudinal study in a triethnic sample of adolescents. Nicotine Tob Res 2003;5:493–506. [13] Tercyak KP. Brief report: social risk factors predict cigarette smoking progression among adolescents with asthma. J Pediatr Psychol 2006; 31:246 –51. [14] Pierce JP, Choi WS, Gilpin EA, et al. Validation of susceptibility as a predictor of which adolescents take up smoking in the United States. Health Psychol 1996;15:355– 61. [15] Pierce JP, Farkas AJ, Evans N, et al. An improved surveillance measure for adolescent smoking. Tob Control 1995;4:S45– 6. [16] Katon WJ, Lozano P, Russo J, et al. The prevalence of DSM-IV anxiety and depressive disorders in youth with asthma. Under review. [17] Shaffer D, Fisher P, Lucas CP, et al. NIMH Diagnostic Interview Schedule for Children Version IV (NIMH DISC-IV): description, differences from previous versions, and reliability of some common diagnoses. J Am Acad Child Adolesc Psychiatry 2000;39:28 –38. [18] Aschenbach TM. Manual for Childhood Behavioral Checklist. Burlington, VT: University of Vermont, 1991. [19] Verhulst FC, van der Ende J. Agreement between parents’ reports and adolescents’ self-reports of problem behavior. J Child Psychol Psychiatry 1992;33:1011–23. [20] Choi WS, Gilpin EA, Farkas AJ, et al. Determining the probability of future smoking among adolescents. Addiction 2001;96:313–23. [21] Asmussen L, Olson LM, Grant EN, et al. Reliability and validity of the Children’s Health Survey for Asthma. Pediatrics 1999;104:e71. [22] National Committee for Quality Assurance (NCQA). HEDIS 3.0 -Health Plan Employer Data and Information Set. Washington, DC: NCQA, 1997. [23] Fishman PA, Shay DK. Development and estimation of a pediatric chronic disease score using automated pharmacy data. Med Care 1999;37:874 – 83. [24] Patton GC, Carlin JB, Coffey C, et al. Depression, anxiety, and smoking initiation: a prospective study over 3 years. Am J Public Health 1998;88:1518 –22. [25] Goodwin RD, Lewinsohn PM, Seeley JR. Cigarette smoking and panic attacks among young adults in the community: the role of parental smoking and anxiety disorders. Biol Psychiatry 2005;58: 686 –93.
[26] Upadhyaya HP, Deas D, Brady KT, et al. Cigarette smoking and psychiatric comorbidity in children and adolescents. J Am Acad Child Adolesc Psychiatry 2002;41:1294 –305. [27] Isensee B, Wittchen HU, Stein MB, et al. Smoking increases the risk of panic: findings from a prospective community study. Arch Gen Psychiatry 2003;60:692–700. [28] Breslau N, Klein DF. Smoking and panic attacks: an epidemiologic investigation. Arch Gen Psychiatry 1999;56:1141–7. [29] Goodman E, Capitman J. Depressive symptoms and cigarette smoking among teens. Pediatrics 2000;106:748 –55. [30] Brook JS, Schuster E, Zhang C. Cigarette smoking and depressive symptoms: a longitudinal study of adolescents and young adults. Psychol Rep 2004;95:159 – 66. [31] Duncan B, Rees DI. Effect of smoking on depressive symptomatology: a reexamination of data from the National Longitudinal Study of Adolescent Health. Am J Epidemiol 2005;162:461–70. [32] Rakowski W, Wells BL, Lasater TM, et al. Correlates of expected success at health habit change and its role as a predictor in health behavior research. Am J Prev Med 1991;7:89 –94. [33] Laukkanen E, Shemeikka S, Notkola IL, et al. Externalizing and internalizing problems at school as signs of health-damaging behaviour and incipient marginalization. Health Promot Int 2002;17: 139 – 46. [34] Navon L, Fiore B, Anderson H. Asthma and tobacco: double trouble for Wisconsin adolescents. WMJ 2005;104:47–53. [35] Choi WS, Pierce JP, Gilpin EA, et al. Which adolescent experimenters progress to established smoking in the United States. Am J Prev Med 1997;13:385–91. [36] Huang M, Hollis J, Polen M, et al. Stages of smoking acquisition versus susceptibility as predictors of smoking initiation in adolescents in primary care. Addict Behav 2005;30:1183–94. [37] Bush T, Curry SJ, Hollis J, et al. Preteen attitudes about smoking and parental factors associated with favorable attitudes. Am J Health Promot 2005;19:410 –7. [38] Weiss JW, Cen S, Schuster DV, et al. Longitudinal effects of protobacco and anti-tobacco messages on adolescent smoking susceptibility. Nicotine Tob Res 2006;8:455– 65. [39] McAfee T, Ludman E, Grothaus L, et al. Physician tobacco advice to preteens in a smoking-prevention randomized trial: steering clear. J Pediatr Psychol 2005;30:371– 6. [40] Nouwen A, Freeston MH, Labbe R, et al. Psychological factors associated with emergency room visits among asthmatic patients. Behav Modif 1999;23:217–33.
Original article
Anxiety and Depressive Disorders Are Associated with Smoking in Adolescents with Asthma Terry Bush, Ph.D.a,*, Laura Richardson, M.D., M.P.H.b, Wayne Katon, M.D.c, Joan Russo, Ph.D.c, Paula Lozano, M.D., M.P.H.b, Elizabeth McCauley, Ph.D.c, and Malia Olivera a
Center for Health Studies, Group Health Cooperative, Seattle, Washington Department of Pediatrics, University of Washington Medical School, Seattle, Washington c Department of Psychiatry & Behavioral Sciences, University of Washington, Seattle, Washington Manuscript received September 18, 2006; manuscript accepted November 30, 2006 b
Abstract
Purpose: To evaluate the association between mental health indicators (including meeting criteria for one or more DSM-IV [Diagnostic and Statistical Manual of Mental Disorders—fourth edition] anxiety or depressive disorders) and susceptibility to smoking or current smoking among youth with asthma and to evaluate the impact of smoking on asthma symptoms and self-management. Methods: We conducted telephone interviews with a population-based sample of 11- to 17-yearold youth and their parents (n ⫽ 769). Interview content included questions on smoking behaviors, asthma symptoms and treatment, externalizing behavior, and a structured psychiatric interview to assess DSM-IV anxiety and depressive disorders. Results: Five percent of youth were smokers and 10.6% indicated that they were “susceptible to smoking.” Smoking was more common among youth with mental health disorders. Anxiety/ depressive disorders were present in 14.5% of nonsmokers, 19.8% of susceptible nonsmokers, and 37.8% of smokers. After controlling for important covariates, youth with more than one anxiety and depressive disorder were at over twofold increased risk for being a smoker. Similarly, for each one-point increase in externalizing disorder symptoms, youth had a 10% increase in likelihood of being a smoker and a 4% increase in risk for “susceptibility to smoking.” Youth who were smokers reported more asthma symptoms, reduced functioning due to asthma, less use of controller medicines, and more use of rescue medications. Conclusions: Comorbid mental health disorders are associated with increased risk of smoking in youth with asthma. Smoking is associated with increased asthma symptom burden and decreased controller medication use. Interventions for youth with asthma should consider screening for and targeting these behavioral concerns. © 2007 Society for Adolescent Medicine. All rights reserved.
Keywords:
Adolescent; Asthma; Smoking; Mental disorder classification; Susceptibility; Anxiety/depression
With a prevalence of up to 10%, asthma is the most common medical disorder among adolescents [1]. When present, asthma is associated with increased functional impairment and lost days at school and increases the risk for development of anxiety or depressive disorders [2,3]. *Address correspondence to: Dr. Terry Bush, Center for Health Studies, Group Health Cooperative, 1730 Minor Ave, Suite 1600, Seattle, WA 98101-1448. E-mail address: [email protected]
Smoking is also common in adolescents, with prevalence rates estimated at 8.3% among middle school students and 21.8% among high school students [4]. Among youth with asthma, prevalence estimates of smoking have been reported to be similar to or higher than those of the general population [5–7]. Smoking and exposure to secondhand smoke are associated with an increased requirement for asthma medication [8], more emergency room visits, and more hospital admissions [9]. Despite known risks, many youth with asthma continue to experiment with smoking, but factors associated with
1054-139X/07/$ – see front matter © 2007 Society for Adolescent Medicine. All rights reserved. doi:10.1016/j.jadohealth.2006.11.145
426
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
smoking onset are not well understood. Common risk factors for youth smoking in general include anxiety [10] and depressive [11] disorders and parental smoking [5,6,12]. Some studies suggest that adolescents with asthma have reasons for smoking similar to those of the general public [7,13]. Common risk factors for smoking in adolescents with or without asthma include exposure to smoking role models, perceived parental approval of smoking, perceived social value of smoking, and rebelliousness [6]. However, a recent study did not find an association between depressive symptom measures and smoking progression over time in youth with asthma [13]. One limitation of this study is that it did not have information on common mental health comorbidities (such as anxiety and externalizing disorders) and lacked information on Diagnostic and Statistical Manual of Mental Disorders—Fourth edition (DSM-IV) diagnoses. Additionally, prior studies did not have data on asthma severity, symptom burden or treatment, which may be important confounders of the relationship between mental health disorders and smoking in addition to understanding the potential impact of smoking on quality of life for these youth. Another limitation of prior studies among youth with asthma is that they do not assess susceptibility to smoke. This is an important variable to measure because the onset of smoking begins in youth with a lack of a firm resolve to not smoke and progresses through stages of susceptibility to smoking including experimentation and occasional use to regular smoking [12,14]. Although smoking in young teens is rare, susceptibility to smoke is more common and is a strong predictor of future smoking [15]. The goals of this study are to evaluate, among youth with asthma, the association between smoking behaviors and mental health symptoms and disorders (externalolizing, anxiety and depressive) and also to evaluate the association between smoking and asthma symptoms and selfmanagement.
Methods We used data from a large epidemiological study of the prevalence of mental health disorders among adolescents with asthma [16] (unpublished data, 2005). Participants were initially identified from automated pharmacy and registration data of Group Health Cooperative (GHC), a large nonprofit health maintenance organization with over 600,000 enrollees in urban and rural settings in Washington State representing about 10% of the state population. GHC enrollees are largely representative of western Washington with the exception of fewer with higher levels of income. Youth were eligible for the study if they had been enrolled in the health plan for at least 6 months, were between 11 and 17 years of age, and had asthma defined as having at least
one of the following in the past 12 months based on review of automated data: hospitalized with an asthma diagnosis and at least one dispensing of an asthma medication during the same time period; more than one emergency room or urgent care visit with an asthma diagnosis; at least two office visits with an asthma diagnosis and at least one dispensing of an asthma medication; only one visit with an asthma diagnosis but at least two dispensings of asthma medications filled on different days; at least four dispensings of asthma medications; at least one office visit with an asthma diagnosis and another in the past 18 months and at least one dispensing of an asthma medication. Youth with evidence of bipolar disorder or schizophrenia based on International Classification of Diseases—ninth version (ICD-9) codes from automated data were excluded. Parents and eligible youth were contacted by phone and invited to participate in the study. Parents were eligible for the adult survey if they were the parent or legal guardian who had the most knowledge of their child’s health and health care. Interviewers from the Center for Health Studies Survey Department obtained verbal informed consent and conducted separate telephone interviews with parents and youth. Interviewers received over 12 hours of training by survey professionals and the study clinician (E.M.). Interview quality and reliability was assured through ongoing monitoring, analysis, feedback, and corrective instructions if needed. Interviewers were blinded to the hypotheses of this study. Participating youth received a $25 coupon to a local store after completion of the youth interview. The Institutional Review Board of Group Health Cooperative approved all study procedures. Among the 1288 families eligible for the study, 769 parents (59.7%) and 781 youth (60.6%) completed the interview (Figure 1). After excluding parents who refused access to youth medical records, our weighted sample size was 757 parents and 769 youth. Comparisons between survey responders and nonresponders revealed that nonresponders were more likely to live in a large town (p ⬍ .02) but less likely to be on Medicaid (p ⬍ .02), to have made a mental health visit (p ⬍ .01), to have received a depression diagnosis (p ⬍ .01), or to have filled a prescription for antidepressants in the past year (p ⬍ .01). Measures Mental health measures The main variables examined in these analyses were presence of DSM-IV anxiety and depressive disorders based on interviews conducted with youth using the Diagnostic Interview Schedule for Children NIMH DISC-4.0 (C-DISC) [17] and externalizing behaviors assessed during the parent
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
427
Figure 1. Recruitment for the Youth with Asthma Study
interview using the Child Behavior Checklist (CBCL) [18]. The C-DISC has been shown to be a reliable and valid structured interview to diagnose DSM-IV disorders in children and adolescents [17]. We administered the C-DISC with adolescents and the CBCL with parents because research indicates that youth are more accurate at reporting symptoms of an internalizing disorder and parents are more accurate at reporting youth externalizing behaviors [19]. We administered the C-DISC depression modules (major depression, dysthymia) and anxiety modules (panic disorder, separation anxiety, social phobia, generalized anxiety and agoraphobia). Analyses were initially conducted separately for major depression, panic disorder, and anxiety disorders. Because cross-sectional and longitudinal data consistently show a high correlation between anxiety and depression, we also created and evaluated a separate variable describing youth with at least one anxiety or depression diagnosis versus none. Smoking measures We assessed adolescent smoking with three standard and validated questions commonly used in youth surveys: 1) “have you ever tried or experimented with cigarette smoking, even a few puffs?,” 2) “have you smoked at least 100 cigarettes in your life?,” and 3) “think about the last 30 days. On how many of these days did you smoke?” [13,20]. The primary outcome measures for this article were current smoking (defined as having smoked at least 100 cigarettes and having smoked in the past month) [20] and susceptibility to smoking [14]. We assessed susceptibility to smoking for nonsmokers with three standard questions: 1) “Do you think you will try a cigarette soon?,” 2) “If one of your best friends were to offer you a cigarette, would you smoke it?,” and 3) “Do you think you will be smoking one year from now?” The response option for questions 2 and 3 is “defi-
nitely yes,” “probably yes,” “probably not,” or “definitely not.” As defined, a person was considered susceptible to smoke if they thought they would try a cigarette soon or if they answered anything other than ‘definitely not’ to questions 2 and 3 [14]. We calculated the total number of youth who were either susceptible to smoking or who were current smokers, because both increase the risk for future smoking. We also asked youth about household smoking with the following, “Has anyone in the family smoked in the last year, including smoking outside of the house?” Asthma measures The diagnosis of asthma for youth was confirmed with parents during their interview. Youth were asked about the age of onset of their asthma and the number of symptom days in the prior 2 weeks. Youth also completed three scales from the Children’s Health Status-Asthma Teen version (CHSA-T), measuring asthma-specific emotional and physical functioning and activity limitations [21]. We used automated diagnostic, visit, and pharmacy data to measure asthma severity based on the Health Employer Data and Information Set (HEDIS) definition of children/ adolescents who are at high risk of adverse asthmatic events [22]. In this definition, meeting criteria for any of the following four health care utilization variables over a 12month period identified individuals at high medical risk: 1) four or more dispensings of asthma medication; 2) one or more emergency visits with an asthma diagnosis; 3) one or more hospitalizations with an asthma diagnosis; 4) four or more ambulatory visits with an asthma diagnosis. Approximately 66% of our sample met the criteria for four or more dispensings of an asthma medication, suggesting lack of specificity as a measure of severity in our population. We therefore developed a second HEDIS measure with a modified medication criteria (replacing criteria 1 above with the
428
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
requirement of having one or more oral steroid prescriptions for asthma in the past 12 months). We tested both of these measures in separate analyses and found similar results. Table 1 shows results using the original HEDIS measure. We also compared smoking groups on use of rescue versus controller medications. Rescue medications such as albuterol are used as needed for patients who continue to have breakthrough symptoms despite use of controller medications. Controller medicines such as fluticasone propionate are intended to be taken on a regular basis to prevent symptoms. The distinction is important because smoking is associated with resistance to medicines, which may lead to increased use of rescue medicines for relief of symptoms. Demographics Demographic variables were evaluated as covariates in the main analysis. Parents were asked about their race, ethnicity, marital status, and the education and employment status of both parents. Children were asked to confirm their age and gender. To assess socioeconomic status, we used the GEOCODE, a measure of neighborhood median household income based on subject zip code. Rural Urban Commuting Area (RUCA) codes were used to further categorize household location as rural or urban. We also included a variable comparing youth from eastern versus western Washington because of the distinct differences between these two areas of the state on population density, environmental exposures to tobacco, socioeconomic status, and health care access. Because of these differences, we hypothesized that there may also be geographic differences in youth smoking rates. Chronic disease measure Chronic diseases other than asthma may be associated with smoking and mental health concerns. To evaluate for possible confounding, we used a modified form of the Pediatric Chronic Disease Scale (PCDS). The PCDS is a measure of medical comorbidity developed based on automated pharmacy data at Group Health and has been shown to correlate with subsequent health care costs and hospitalization rates [23]. For this study, the PCDS was modified by removing medications that are commonly used for asthma and anxiety or depression.
of Diseases—ninth version (ICD-9) diagnosis of depression or anxiety. In conducting this propensity analysis, we first used logistic regression to estimate a response probability for each survey respondent (response propensity score). Then, we calculated weights that were inversely proportional to estimated probability of response rescaled to sum to the observed sample size (i.e., the number of survey respondents). In weighted analyses, individuals with a low probability of response were given higher weights to represent the larger number of nonrespondents with similar characteristics. For this report, all regression analyses were conducted with propensity score weighting. First, we conducted univariate analyses comparing all demographic, asthma, and mental health variables among the three smoking groups. For these analyses, we used chi-square tests with corrections for continuity and analyses of variance (ANOVAs) for categorical and continuous variables, respectively. In the event of a significant global group difference, we conducted two planned tests (either ANOVA or chi-square) to determine which set of group differences was responsible for the statistical significance: nonsmokers versus smokers or nonsmokers versus susceptible nonsmokers. To determine which factors were independently and significantly associated with susceptibility to smoking and current smoking in comparison to nonsmoking, we conducted propensity weighted logistic regression analyses. Two sets of regression analyses were run. One compared the smoking with the nonsmoking group and a second compared the susceptible to smoking with the nonsmoking group. All factors that were statistically significant in bivariate analyses (p ⬍ .05) were initially entered into the models. Then, individually, the least significant variable (the variable with the largest p-value that was greater than .05) was eliminated and the models were refit. This was continued until the models contained only statistically significant (p ⬍ .05) factors. We then estimated the odds ratios (OR) and 95% confidence intervals (CI) for these variables. To make sure that no other significant variables were omitted from the models, the significance of all variables added to the model were assessed.
Analysis
Results
Using de-identified data, we conducted a nonresponder analysis to determine if survey respondents differed significantly from nonrespondents. We used propensity analysis to estimate the probability of being a respondent as a function of age, gender, RUCA code (rural versus urban area), being on Medicaid or state-funded insurance because of low income, PCDS, number of primary care visits, number of asthma-related emergency room visits and hospitalizations, oral steroid prescription, number of specialty mental health visits, any prescription for antidepressant medications or anti-anxiety medications, and an International Classification
Table 1 presents study variables for the total sample completing the interview as well as study variables stratified by youth smoking status. Among participating parents, 46% were female, 92.1% were educated beyond high school, 93.8% were employed, 72.9% were married, and 80.2% were white. Among participating adolescents (mean age ⫽ 14 years), 46.6% were female, 47.3% reported exposure to household smokers, 71.2% were on an asthma controller medication, and 16.2% met DSM-IV criteria for one or more anxiety or depressive disorders. As shown in Table 1, 15.6% of youth were either current smokers or susceptible
Table 1 Characteristics of adolescents with asthma by smoking status Current smoker n ⫽ 38 (5.0%)
Nonsmoker susceptible n ⫽ 82 (10.6%)
Nonsmoker not susceptible n ⫽ 649 (84.4%)
Overall 2 or F test
Current vs. Nonsmoker
Susceptible vs. Nonsmoker
60 (7.9%) 710 (93.8%) 551 (72.9%) 606 (80.2%) 104 (13.5%)
7 (19.4%) 31 (86.1%) 15 (41.7%) 27 (73.0%) 3 (7.9%)
9 (11.1%) 77 (95.1%) 62 (76.5%) 62 (76.5%) 12 (14.8%)
44 (6.9%) 602 (94.1%) 474 (74.2%) 517 (81.0%) 89 (13.7%)
p ⬍ .01 ns p ⬍ .001 ns ns
p ⬍ .05 ND p ⬍ .001 ND ND
ns ND ns ND ND
50,460 ⫾ 17,015
49,385 ⫾ 13,884
51,962 ⫾ 20,418
50,322 ⫾ 16,696
ns
ND
ND
359 (46.6%) 14.0 ⫾ 1.9
28 (73.7%) 16.1 ⫾ .8
33 (40.2%) 14.4 ⫾ 1.6
298 (45.8%) 13.8 ⫾ 1.9
p ⬍ .01 p ⬍ .001
p ⬍ .001 p ⬍ .001
ns p ⬍ .05
213 (27.7%) 234 (30.4%) 323 (41.9%) 362 (47.3%)
0 (0.0%) 1 (2.6%) 37 (97.4%) 33 (86.8%)
11 (13.4%) 34 (41.5%) 37 (45.1%) 51 (63.0%)
202 (31.1%) 199 (30.6%) 249 (38.3%) 285 (43.0%)
p ⬍ .001
p ⬍ .001
p ⬍ .01
p ⬍ .001
p ⬍ .001
p ⬍ .001
ns
ND
6.7 ⫾ 4.3
7.8 ⫾ 5.1
7.3 ⫾ 4.4
6.6 ⫾ 4.2
3.8 ⫾ 4.0 83.7 ⫾ 17.1 83.7 ⫾ 14.6 73.7 ⫾ 15.6 .40 ⫾ .70
6.3 ⫾ 5.2 76.8 ⫾ 19.2 79.0 ⫾ 16.1 65.4 ⫾ 14.7 .64 ⫾ .96
4.4 ⫾ 4.6 83.6 ⫾ 18.7 83.3 ⫾ 14.4 71.9 ⫾ 16.2 .38 ⫾ .63
3.6 ⫾ 3.8 84.1 ⫾ 16.7 84.2 ⫾ 14.5 74.4 ⫾ 15.4 .39 ⫾ .69
p ⬍ .001 p ⬍ .05 ns p ⬍ .001 ns
p ⬍ .001 p ⬍ .01 ND p ⬍ .01 ND
ND ns p ⬍ .07 ND ns ND
614 ⫾ 960
859 ⫾ 1294
614 ⫾ 806
600 ⫾ 954
ns
ND
ND
p ⬍ .05 (df ⫽ 4)
p ⬍ .05
ns
26 (3.4%) 197 (25.6%) 548 (71.2%)
0 (0.0%) 16 (41.0%) 23 (59.0%)
4 (4.9%) 26 (31.7%) 52 (63.4%)
22 (3.4%) 155 (23.8%) 473 (72.9%)
9.2 ⫾ 8.2 56 (8.0%) 19 (2.5%) 104 (13.6%)
17.4 ⫾ 11.8 6 (20.7%) 2 (5.6%) 11 (29.7%)
11.5 ⫾ 9.3 10 (13.3%) 3 (3.7%) 14 (17.3%)
8.5 ⫾ 7.6 40 (6.7%) 14 (2.2%) 79 (12.2%)
p ⬍ .001 p ⬍ .01 ns p ⬍ .01
p ⬍ .001 p ⬍ .01 ND p ⬍ .01
p ⬍ .001 ns ND p ⬍ .001
124 (16.2%)
14 (37.8%)
16 (19.8%)
94 (14.5%)
p ⬍ .001
p ⬍ .001
ns
429
ND ⫽ not determined because the overall test was not significant; HEDIS ⫽ Health plan Employer Data and Information Set; PCDS ⫽ pediatric chronic disease score; MH ⫽ mental health diagnosis. a Children’s Health Status-Asthma Teen version (CHSA-T). b Modified HEDIS revealed similar results. c Parent rated (Child Behavior Checklist).
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
Parent demographics Highest education level high school or less, either parent Employed full or part time, either parent % Married % Caucasian, participating parent % Medicaid GEOCODE (mean household income based on zip code) Youth demographics % female Mean age ⫾ SD Age groups (years) % 11–12 % 13–14 % 15–17 % Households with smoker (past year) Asthma variables Age onset with asthma Asthma severity: Symptom days past 2 weeks Health status–Activitiesa Health status–Emotional Health status–Physical HEDIS Measure (0–4)b Medical comorbidity score (PCDS w/o asthma, w/o a MH diagnosis) Asthma treatment intensity No medications Albuterol only At least 1 controller Mental health Externalizing scalec % With major depression % With panic disorder % With any anxiety disorder % With 1⫹ anxiety or depression disorders
Total n ⫽ 769 (100%)
430
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
to smoking in the future. Although current smoking was present in only 5%, it was more common among older adolescents (11.5% among 15–17-year-olds and less than 1% [n ⫽ 1] in 11–14-year-olds). Smokers significantly differed from nonsmokers on many important factors. In comparison with nonsmokers, smokers were significantly older, more likely to be female, less likely to come from homes where parents were married and educated beyond high school, and more likely to report exposure to household smoking. Smokers also reported more asthma symptom days in the past 2 weeks and more functional limitations (CHSA-T activities and physical health scales). Smokers were more likely to use rescue medications such as albuterol and less likely to use controller medications despite having more asthma symptom days per week. Each of the mental health disorder categories examined were more common among smokers: smokers were more likely to meet DSM-IV diagnostic criteria for a depressive or anxiety disorder, and had higher parent-rated externalizing disorder symptoms. Compared with nonsusceptible, nonsmoking youth, susceptible nonsmoking youth were older, more likely to have a family member that smoked, had higher mean externalizing symptom scores, and were more likely to have an anxiety disorder. The multivariate logistic model controlling for significant covariates (Table 2) showed that, in comparison to nonsusceptible nonsmoking youth, those who smoked were significantly more likely to be older (OR 2.90, 95% CI 1.98 – 4.26) and female (OR 2.40, 95% CI 1.01–5.72) and to have one or more family members who smoked (OR 6.59, 95% CI 2.37–18.31). They had significantly higher parentrated externalizing behavior symptom scores (OR 1.10, 95% CI 1.05–1.14) and were more likely to meet DSM-IV criteria for one or more anxiety or depressive disorders (OR 2.58, 95% CI 1.06 – 6.28). Compared with nonsusceptible nonsmoking youth, youth who were susceptible to smoking (Table 3) were significantly older (OR 1.18, 95% CI 1.04 –1.34), more likely to have one or more family members who smoked (OR 1.92, 95% CI 1.17–3.14), and had higher externalizing scale scores (OR 1.04, 95% CI 1.01–1.07). After controlling for
Table 2 Weighted logistic regression for adolescent smokers versus the nonsmokers Characteristic
Odds ratio
95% CI
p
Youth age Household smoking Female (youth) CBCL externalizing behavior DSM-IV anxiety-depressive disorder
2.904 6.591 2.399 1.097
(1.978–4.262) (2.372–18.314) (1.006–5.721) (1.053–1.144)
.000 .000 .048 .000
2.577
(1.057–6.284)
.037
CI ⫽ confidence interval; CBCL ⫽ Child Behavior Checklist; DSM-IV ⫽ Diagnostic and Statistical Manual– 4th edition.
Table 3 Weighted logistic regression for the susceptible versus nonsusceptible nonsmokers Characteristic
Odds ratio
95% CI
p
Youth age Household smoking CBCL externalizing behavior
1.179 1.920 1.037
(1.040–1.337) (1.172–3.143) (1.010–1.065)
.010 .010 .008
CI ⫽ confidence interval; CBCL ⫽ Child Behavior Checklist.
covariates, anxiety disorders were no longer significantly associated with susceptibility to smoke.
Discussion In this population-based sample of youth with asthma, we found that youth with comorbid anxiety and depressive disorders or elevated externalizing symptoms were more likely to be smokers. These associations persisted even after controlling for potentially important demographic and clinical covariates. Adolescents with asthma who smoked were more than twice as likely to have major depression (20.7% vs. 6.7%) and one or more anxiety disorders (29.7% vs. 12.2%) compared with nonsmokers. This finding is consistent with studies of youth respondents from the community where anxiety and depressive disorders have been shown to be associated with increased risk for smoking [10,11]. Our findings are also consistent with increasing evidence in cross-sectional and longitudinal studies of a link between panic attacks, depression and anxiety disorders, and smoking [10,11,24 –26]. The results of this study are in contrast with a recent study of youth with asthma which found that depressive symptoms or sadness did not increase the risk for continued smoking [13]. One reason our study results may differ is that we looked at the association cross-sectionally rather than longitudinally. We also had more specific criteria for delineating mental health disorders, including the use of structured psychiatric interviews and parental ratings of externalizing symptoms. Because of the cross-sectional nature of our data, we cannot determine the direction of association between smoking and anxiety or depressive disorders. Prior studies suggest that the relationship between smoking and mental disorders may be bidirectional [10]. Analysis of longitudinal data indicates that depression or panic [10,27] is associated with a higher risk for onset of smoking and the reverse, that smoking is associated with a higher risk for development of panic [10,27,28] and major depression [10,29,30]. It has also been suggested that the association between mental health factors and smoking may be due to underlying causal factors such as adverse childhood experiences, which are potential antecedents to both depression/ anxiety and smoking [31].
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
To our knowledge, this is the first study to examine externalizing disorder symptoms as a predictor of smoking among asthmatics. However, our finding of an increased risk for smoking in youth with asthma with a high level of externalizing symptoms is consistent with studies in the general population that have shown that youth with externalizing disorders are at increased risk for smoking [32,33]. It is important to note that 87% of the smokers reported at least one family member who smoked. In addition to the risk for the development of smoking, exposure to secondhand smoke in the home increases the risk for development of asthma [34] and may be a contributing factor to the increase in asthma severity we observed in youth smokers. Youth who smoked reported more symptom days, more functional impairment, and more overall asthma symptoms. Although it is unclear if this is a result of patient or health care factors, smokers also reported less use of medications to control asthma and higher use of rescue medications (e.g., albuterol) compared with nonsmokers. Even when youth are using controllers, studies suggest that exposure to tobacco smoke may reduce the effectiveness of asthma medications [8]. One potential limitation of this study is that the prevalence of smoking in our population was low at 5%. This may be due to the young age of our sample. When examining the older age group, the prevalence of smoking was 11.5%. Additionally, it may be due to regional differences in smoking prevalence. The prevalence of smoking is lower in the Pacific Northwest, where this study was conducted, than in other regions of the country. Although smoking was relatively infrequent in our population, twice as many were susceptible to smoking. Studies repeatedly show that susceptibility to smoking and smoking during childhood are strong predictors of future smoking [14,15,35,36]. Given the high risk, these susceptible youth might be a good target for primary prevention. One potential intervention target is media images of smoking. Youth of this age are particularly vulnerable to peer and media influences and begin to formulate positive attitudes about smoking without realizing the addictive properties of nicotine [37]. In a recent longitudinal study, nonsusceptible youth who recalled pro-tobacco messages or role models in print or visual forms (stores, magazines, TV, movies) were at increased risk for subsequent smoking susceptibility [38]. In addition to the low prevalence of smoking and the cross-sectional design, another limitation of this study is the use of a limited number of measures to assess smoking. We recognize there are other ways of measuring and defining smoking and that our measure relies on self-report. Another potential limitation to this study is that the C-DISC was administered only to the youth. It is possible that parentreported anxiety and depression from the DISC may have been different from youth self-report. However, there is controversy in the scientific field about how to include parent data when there is discordance between parent and child data. Research indicates that youth are more accurate
431
at reporting internalizing symptoms than their parents [19]. Strengths of this study include the large population-based sample, the careful assessment of DSM-IV anxiety and depression disorders, and the availability of measures of asthma symptoms and treatment. The results of our study have clinical and public health significance. We have identified important risk factors for smoking and susceptibility to smoke in youth with asthma and provide evidence of the impact of smoking on management of asthma. Clinicians are in a unique position to identify and intervene with youth to encourage quitting or reinforce not smoking [39]. Primary care physicians often screen children/adolescents with asthma for smoking. Our data suggest that it is important to also screen these children/adolescents for anxiety and depressive disorders as well as externalizing behaviors, because of the increased association of these mental health problems with smoking. Efforts to educate the public about the association between smoking and mental disorders are also needed because each has been associated with adverse outcomes among those with asthma [40]. Our results confirm the importance of addressing risk factors for smoking in youth with asthma and developing early prevention and cessation interventions, including efforts to reduce exposure to secondhand smoke. Acknowledgments This work was supported by National Institute of Mental Health grant #MH 67587 (Principal Investigator: Wayne Katon, MD). References [1] Weitzman M, Gortmaker SL, Sobol AM, et al. Recent trends in the prevalence and severity of childhood asthma. JAMA 1992;268: 2673–7. [2] Katon WJ, Richardson L, Lozano P, et al. The relationship of asthma and anxiety disorders. Psychosom Med 2004;66:349 –55. [3] Goodwin RD, Fergusson DM, Horwood LJ. Asthma and depressive and anxiety disorders among young persons in the community. Psychol Med 2004;34:1465–74. [4] Centers for Disease Control and Prevention (CDC). Tobacco use, access, and exposure to tobacco in media among middle and high school students—United States, 2004. MMWR 2005;54:297–301. [5] Otten R, Engels RC, van den Eijnden RJ. Parental smoking and smoking behavior in asthmatic and nonasthmatic adolescents. J Asthma 2005;42:349 –55. [6] Zbikowski SM, Klesges RC, Robinson LA, et al. Risk factors for smoking among adolescents with asthma. J Adolesc Health 2002;30: 279 – 87. [7] Precht DH, Keiding L, Nielsen GA, et al. Smoking among upper secondary pupils with asthma: reasons for their smoking behavior: a population-based study. J Adolesc Health 2006;39:141–3. [8] Strachan DP, Cook DG. Health effects of passive smoking: parental smoking and childhood asthma: longitudinal and case-control studies. Thorax 1998;53:204 –12. [9] Troisi RJ, Speizer FE, Rosner B, et al. Cigarette smoking and incidence of chronic bronchitis and asthma in women. Chest 1995;108: 1557– 61.
432
T. Bush et al. / Journal of Adolescent Health 40 (2007) 425– 432
[10] Johnson JG, Cohen P, Pine DS, et al. Association between cigarette smoking and anxiety disorders during adolescence and early adulthood. JAMA 2000;284:2348 –51. [11] Covey LS, Tam D. Depressive mood, the single-parent home, and adolescent cigarette smoking. Am J Public Health 1990;80:1330 –3. [12] Gritz ER, Prokhorov AV, Hudmon KS, et al. Predictors of susceptibility to smoking and ever smoking: a longitudinal study in a triethnic sample of adolescents. Nicotine Tob Res 2003;5:493–506. [13] Tercyak KP. Brief report: social risk factors predict cigarette smoking progression among adolescents with asthma. J Pediatr Psychol 2006; 31:246 –51. [14] Pierce JP, Choi WS, Gilpin EA, et al. Validation of susceptibility as a predictor of which adolescents take up smoking in the United States. Health Psychol 1996;15:355– 61. [15] Pierce JP, Farkas AJ, Evans N, et al. An improved surveillance measure for adolescent smoking. Tob Control 1995;4:S45– 6. [16] Katon WJ, Lozano P, Russo J, et al. The prevalence of DSM-IV anxiety and depressive disorders in youth with asthma. Under review. [17] Shaffer D, Fisher P, Lucas CP, et al. NIMH Diagnostic Interview Schedule for Children Version IV (NIMH DISC-IV): description, differences from previous versions, and reliability of some common diagnoses. J Am Acad Child Adolesc Psychiatry 2000;39:28 –38. [18] Aschenbach TM. Manual for Childhood Behavioral Checklist. Burlington, VT: University of Vermont, 1991. [19] Verhulst FC, van der Ende J. Agreement between parents’ reports and adolescents’ self-reports of problem behavior. J Child Psychol Psychiatry 1992;33:1011–23. [20] Choi WS, Gilpin EA, Farkas AJ, et al. Determining the probability of future smoking among adolescents. Addiction 2001;96:313–23. [21] Asmussen L, Olson LM, Grant EN, et al. Reliability and validity of the Children’s Health Survey for Asthma. Pediatrics 1999;104:e71. [22] National Committee for Quality Assurance (NCQA). HEDIS 3.0 -Health Plan Employer Data and Information Set. Washington, DC: NCQA, 1997. [23] Fishman PA, Shay DK. Development and estimation of a pediatric chronic disease score using automated pharmacy data. Med Care 1999;37:874 – 83. [24] Patton GC, Carlin JB, Coffey C, et al. Depression, anxiety, and smoking initiation: a prospective study over 3 years. Am J Public Health 1998;88:1518 –22. [25] Goodwin RD, Lewinsohn PM, Seeley JR. Cigarette smoking and panic attacks among young adults in the community: the role of parental smoking and anxiety disorders. Biol Psychiatry 2005;58: 686 –93.
[26] Upadhyaya HP, Deas D, Brady KT, et al. Cigarette smoking and psychiatric comorbidity in children and adolescents. J Am Acad Child Adolesc Psychiatry 2002;41:1294 –305. [27] Isensee B, Wittchen HU, Stein MB, et al. Smoking increases the risk of panic: findings from a prospective community study. Arch Gen Psychiatry 2003;60:692–700. [28] Breslau N, Klein DF. Smoking and panic attacks: an epidemiologic investigation. Arch Gen Psychiatry 1999;56:1141–7. [29] Goodman E, Capitman J. Depressive symptoms and cigarette smoking among teens. Pediatrics 2000;106:748 –55. [30] Brook JS, Schuster E, Zhang C. Cigarette smoking and depressive symptoms: a longitudinal study of adolescents and young adults. Psychol Rep 2004;95:159 – 66. [31] Duncan B, Rees DI. Effect of smoking on depressive symptomatology: a reexamination of data from the National Longitudinal Study of Adolescent Health. Am J Epidemiol 2005;162:461–70. [32] Rakowski W, Wells BL, Lasater TM, et al. Correlates of expected success at health habit change and its role as a predictor in health behavior research. Am J Prev Med 1991;7:89 –94. [33] Laukkanen E, Shemeikka S, Notkola IL, et al. Externalizing and internalizing problems at school as signs of health-damaging behaviour and incipient marginalization. Health Promot Int 2002;17: 139 – 46. [34] Navon L, Fiore B, Anderson H. Asthma and tobacco: double trouble for Wisconsin adolescents. WMJ 2005;104:47–53. [35] Choi WS, Pierce JP, Gilpin EA, et al. Which adolescent experimenters progress to established smoking in the United States. Am J Prev Med 1997;13:385–91. [36] Huang M, Hollis J, Polen M, et al. Stages of smoking acquisition versus susceptibility as predictors of smoking initiation in adolescents in primary care. Addict Behav 2005;30:1183–94. [37] Bush T, Curry SJ, Hollis J, et al. Preteen attitudes about smoking and parental factors associated with favorable attitudes. Am J Health Promot 2005;19:410 –7. [38] Weiss JW, Cen S, Schuster DV, et al. Longitudinal effects of protobacco and anti-tobacco messages on adolescent smoking susceptibility. Nicotine Tob Res 2006;8:455– 65. [39] McAfee T, Ludman E, Grothaus L, et al. Physician tobacco advice to preteens in a smoking-prevention randomized trial: steering clear. J Pediatr Psychol 2005;30:371– 6. [40] Nouwen A, Freeston MH, Labbe R, et al. Psychological factors associated with emergency room visits among asthmatic patients. Behav Modif 1999;23:217–33.