Abstract
Over 40% of all US smokers have comorbid alcohol, drug, or mental disorders. Using data from the 2000–2001 Healthcare for Communities survey, we conducted multivariate logistic regressions to examine these individuals' sensitivity to cigarette prices. We found that a 10% increase in cigarette prices was associated with 18.2% less smoking participation among individuals with alcohol, drug, or mental disorders, except those with alcohol dependence. Increasing cigarette taxes could be effective in reducing smoking among individuals with alcohol, drug, or mental disorders.
Over 40% of all US smokers have comorbid alcohol, drug, or mental disorders.1 Individuals in this group have a higher smoking prevalence1–6 and are more likely to be heavy smokers1 than are those without such disorders. Determining this group's sensitivity to cigarette prices1 is important for tobacco control policy. The only prior study of this issue,7 which used data from 1991, found that individuals with mental illness during the past year had a price elasticity of cigarette smoking participation of −0.533, compared with −0.731 for the overall population. However, this study did not examine individuals with comorbid alcohol or drug disorders. We examined smoking participation and sensitivity to cigarette prices among individuals with comorbid alcohol, drug, or mental disorders.
METHODS
We analyzed the largest sample (7909 individuals) with nationally representative weights from the 2000–2001 Healthcare for Communities household survey, a nationwide survey that specifically focused on alcohol, drug, or mental disorders.8 Respondents who answered “don't know” or “refused” to any of the analytical variables, or who were missing data for any of these items, were deleted from the study, leaving a final sample of 7530 individuals.
Variables
Cigarette use was identified by answers to the question, “Do you currently smoke or chew tobacco?” Annual average state cigarette prices (including generics)9 were logarithmically transformed and assigned to individual respondents by state of residence and year of survey response.
Comorbid alcohol, drug, or mental disorders were constructed from self-report of past-year specific disorders or symptoms. These disorders included psychotic disorder (lifetime diagnosis of schizophrenia or hospitalization for psychosis); generalized anxiety disorder, major depressive disorder, or dysthymia (Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition10 classifications); panic disorder (any past-year associated symptoms); alcohol dependence (Alcohol Use Disorders Identification Test score ≥ 8)11; binge drinking (≥ 6 drinks consumed per occasion); drug dependence (presence of dependence symptoms, or psychological or emotional problems with substance use); and drug abuse (any substance use or prescription drug use that did not follow prescribed directions). Full descriptions of these measures and the survey response rate are available elsewhere.12
Analysis
We calculated weighted prevalence rates for alcohol, drug, and mental disorders and current smoking. We compared the smoking prevalence rates of those with and without such disorders using normalized, weighted, bivariate logistic regression. We conducted normalized, weighted, multivariate logistic regressions, adjusting for covariates (including sociodemographics; specific past-year alcohol, drug, or mental disorder; obesity; and physical activity) to examine the likelihood of being a current smoker among all individuals (full sample) and among individuals with a past-year alcohol, drug, or mental disorder. Adjusted odds ratios and their 95% confidence intervals were estimated for each covariate. All statistical analyses were performed with SAS version 9 (SAS Institute, Cary, North Carolina). Statistical significance was defined at α < .05.
RESULTS
Table 1 shows the weighted prevalence rates of alcohol, drug, or mental disorders and current smoking. Of the 7530 individuals, 23.0% had an alcohol, drug, or mental disorder in the past year. The prevalence of specific alcohol, drug, or mental disorders was lowest for psychosis (1.0%; lifetime) and highest for major depressive disorder (9.4%). The prevalence rates of current smoking were 24.3% for the full sample; 43.8% for the alcohol, drug, or mental disorder sample; and 18.5% for those without alcohol, drug, or mental disorders. Smoking prevalence rates varied by disorder, ranging from 40.5% for those with drug abuse to 61.1% for those with alcohol dependence. Among current smokers, 41.4% had some alcohol, drug, or mental disorder in the past year.
TABLE 1.
Weighted Prevalence Rates of Alcohol, Drug, or Mental (ADM) Disorders and Current Smoking: Healthcare for Communities Survey, 2000–2001
| Characteristic | % of Participants | Current Smoking Prevalence, % |
| Total sample | 100.0 | 24.3 |
| Any ADM disorder | ||
| Yes | 23.0 | 43.8** |
| No | 77.0 | 18.5 |
| Psychotic disorder (lifetime) | ||
| Yes | 1.0 | 44.9** |
| No | 99.0 | 21.1 |
| Generalized anxiety disorder | ||
| Yes | 3.8 | 43.9** |
| No | 96.2 | 23.5 |
| Major depression disorder | ||
| Yes | 9.4 | 47.1** |
| No | 90.6 | 22.0 |
| Dysthymia | ||
| Yes | 4.1 | 44.5** |
| No | 95.9 | 23.5 |
| Panic disorder | ||
| Yes | 3.1 | 46.0** |
| No | 96.9 | 23.6 |
| Alcohol dependence | ||
| Yes | 5.1 | 61.1** |
| No | 94.9 | 22.3 |
| Binge drinking | ||
| Yes | 3.5 | 49.0** |
| No | 96.5 | 23.4 |
| Drug dependence | ||
| Yes | 1.9 | 43.7** |
| No | 98.1 | 24.0 |
| Drug abuse | ||
| Yes | 8.3 | 40.5** |
| No | 91.7 | 22.9 |
Note. Significance testing was based on normalized weighted analyses. Percentages were nationally weighted to approximate the US population as determined from the 2000–2001 Healthcare for Communities survey. The study's total sample size was 7530.
**P < .001.
Cigarette prices had a significant negative effect on smoking participation (price elasticity = −1.82; P = .005) among the alcohol, drug, or mental disorder sample, but not among the full sample (Table 2). In both samples, alcohol dependence and depression were significantly associated with higher smoking participation whereas binge drinking was significantly associated with lower smoking participation (Table 2).We conducted prespecified subanalyses for the alcohol, drug, or mental disorder sample. With stepwise addition of specific alcohol, drug, and mental disorder variables into the model, we found that control for alcohol dependence alone was responsible for uncovering the relationship between cigarette prices and smoking participation (price elasticity = −1.64; P = .011). Excluding individuals with alcohol dependence resulted in a similar significant negative price effect on smoking participation (price elasticity = −1.83; P = .013). When only those with alcohol dependence were analyzed, we found no significant price effect on smoking participation.
TABLE 2.
Adjusted Odds Ratios (AORs) and Price Elasticity Estimates From Multivariate Logistic Regression Models Predicting Smoking Participation: Healthcare for Communities Survey, 2000–2001
| Variable | Full Sample, AOR (95% CI) | ADM Sample, AOR (95% CI) |
| Female sex (reference = male) | 0.48* (0.42, 0.55) | 0.59* (0.47, 0.74) |
| Age, y (reference = 18–24) | ||
| 25–34 | 2.93* (2.22, 3.87) | 2.42* (1.63, 3.59) |
| 35–44 | 4.06* (3.09, 5.33) | 3.61* (2.43, 5.37) |
| 45–54 | 3.54* (2.67, 4.69) | 2.10* (1.37, 3.21) |
| 55–64 | 2.79* (2.07, 3.76) | 2.28* (1.42, 3.64) |
| ≥ 65 | 0.90 (0.63, 1.29) | 0.74 (0.42, 1.31) |
| Ethnicity (reference = White) | ||
| Black | 0.68* (0.56, 0.82) | 0.81 (0.58, 1.12) |
| Hispanic | 0.58* (0.44, 0.75) | 0.40* (0.27, 0.61) |
| Other race | 0.95 (0.64, 1.40) | 1.34 (0.70, 2.58) |
| Region (reference = Northeast) | ||
| Midwest | 0.92 (0.74, 1.15) | 0.93 (0.63, 1.37) |
| South | 0.87 (0.69, 1.09) | 0.81 (0.55, 1.20) |
| West | 0.53* (0.44, 0.65) | 1.27 (0.92, 1.75) |
| Education (reference = less than high school) | ||
| College graduate or higher | 0.17* (0.13, 0.22) | 0.12* (0.08, 0.19) |
| High school graduate or some college | 0.58* (0.48, 0.70) | 0.60* (0.43, 0.83) |
| Not born in US (reference = born in US) | 0.26* (0.19, 0.35) | 0.20* (0.13, 0.33) |
| Income, $ (reference = < 25 000) | ||
| 25 000–50 000 | 1.13 (0.96, 1.33) | 0.85 (0.65, 1.11) |
| 50 000–75 000 | 0.91 (0.75, 1.10) | 1.07 (0.76, 1.52) |
| ≥ 75 000 | 0.81* (0.65, 0.99) | 0.50* (0.34, 0.72) |
| Marital status (reference = married) | ||
| Has partner | 2.42* (1.88, 3.12) | 1.11 (0.75, 1.64) |
| Single | 1.27* (1.10, 1.46) | 1.34* (1.05, 1.72) |
| Unemployed (reference = employed) | 0.91 (0.77, 1.08) | 0.84 (0.64, 1.11) |
| Body mass index (reference = < 25) | ||
| 25–30 | 0.78* (0.68, 0.90) | 0.75* (0.58, 0.96) |
| ≥ 30 | 0.72* (0.61, 0.84) | 0.60* (0.46, 0.78) |
| Activity (reference = none) | ||
| Some or fairly active | 0.67* (0.47, 0.98) | 0.69 (0.41, 1.16) |
| Quite active | 0.73 (0.50, 1.08) | 0.46* (0.26, 0.84) |
| Very active or more | 0.68* (0.46, 0.99) | 0.53* (0.30, 0.95) |
| Insurance (reference = private) | ||
| Medicaid | 1.75* (1.24, 2.46) | 0.85 (0.52, 1.40) |
| Medicare | 1.46* (1.11, 1.92) | 1.13 (0.74, 1.73) |
| Not Insured | 1.72* (1.41, 2.10) | 1.72* (1.27, 2.34) |
| Other | 1.08 (0.70, 1.66) | 0.58 (0.30, 1.12) |
| ADM disorder (reference = none) | ||
| Psychosis (lifetime) | 0.81 (0.46, 1.41) | 0.90 (0.54, 1.50) |
| Generalized anxiety disorder | 1.34 (0.98, 1.83) | 1.06 (0.79, 1.41) |
| Depression | 2.45* (2.00, 3.01) | 1.71* (1.35, 2.16) |
| Dysthymia | 1.03 (0.76, 1.41) | 0.82 (0.61, 1.11) |
| Panic disorder | 0.99 (0.71, 1.40) | 0.87 (0.64, 1.19) |
| Alcohol dependence | 4.45* (3.37, 5.88) | 3.61* (2.72, 4.80) |
| Binge drinking | 0.71* (0.51, 1.00) | 0.58* (0.42, 0.80) |
| Drug dependence | 0.80 (0.50, 1.29) | 0.88 (0.57, 1.37) |
| Drug abuse | 1.57* (1.26, 1.97) | 0.99 (0.77, 1.29) |
| Price elasticity estimate (95% CI) | −0.40 (−1.14, 0.34) | −1.82* (−3.10, −0.54) |
Note. ADM = alcohol, drug, or mental; CI = confidence interval. For the full sample, n = 7530; for the ADM sample, n = 1206.
*P < .05.
DISCUSSION
Consistent with previous studies,1 our results showed that approximately 40% of current smokers had comorbid alcohol, drug, or mental disorders. We found that smoking participation for individuals with the specified alcohol, drug, or mental disorders was significantly sensitive to cigarette prices: a 10% price increase would result in an 18.2% decline in smoking participation. However, our cross-sectional data analyses could not determine if this relationship was causal. As our data did not contain information on quantities of tobacco products consumed, we could not identify the relationship between cigarette prices and consumption. Therefore, our estimates likely underestimate the overall price effect on smoking behavior. The latest data available for us were from 2001 to 2002. Timely data sets and further analyses on this topic are needed to capture more recent smoking patterns among these individuals.
Our study suggests that increasing cigarette prices through tobacco taxation could be an effective policy tool for reducing smoking among individuals with alcohol, drug, or mental disorders, except among those with alcohol dependence. Whether recent cigarette price increases have reduced smoking among individuals with such disorders, and whether the identified association is causal, are questions requiring further study.
Acknowledgments
This work was supported by funding from the Robert Wood Johnson Foundation's Substance Abuse Policy Research Program (No. 61104); a research seed grant from the Jonsson Cancer Center Foundation at the University of California, Los Angeles (UCLA); and the UCLA-RAND NIMH Partnered Research Center for Quality Care (NIMH P30 MH082760).
We thank Teh-wei Hu and Theodore Keeler for comments on prior versions of this brief.
Human Participant Protection
This study was approved by the institutional review board at the University of California, Los Angeles, with a waiver from obtaining informed consent from participants.
References
- 1.Lasser K, Boyd JW, Woolhandler S, Himmelstein DU, McCormick D, Bor DH. Smoking and mental illness: a population-based prevalence study. JAMA 2000;284(20):2606–2610 [DOI] [PubMed] [Google Scholar]
- 2.Breslau N. Psychiatric comorbidity of smoking and nicotine dependence. Behav Genet 1995;25(2):95–101 [DOI] [PubMed] [Google Scholar]
- 3.Glassman AH, Helzer JE, Covey LS, et al. Smoking, smoking cessation, and major depression. JAMA 1990;264(12):1546–1549 [PubMed] [Google Scholar]
- 4.Goff DC, Henderson DC, Amico E. Cigarette smoking in schizophrenia: relationship to psychopathology and medication side effects. Am J Psychiatry 1992;149(9):1189–1194 [DOI] [PubMed] [Google Scholar]
- 5.Gonzalez-Pinto A, Gutierrez M, Ezcurra J, et al. Tobacco smoking and bipolar disorder. J Clin Psychiatry 1998;59(5):225–228 [DOI] [PubMed] [Google Scholar]
- 6.Kelly C, McCreadie RG. Smoking habits, current symptoms, and premorbid characteristics of schizophrenic patients in Nithsdale, Scotland. Am J Psychiatry 1999;156(11):1751–1757 [DOI] [PubMed] [Google Scholar]
- 7.Saffer H, Dave D. Mental illness and the demand for alcohol, cocaine, and cigarettes. Econ Inq 2005;43(2):229–246 [Google Scholar]
- 8.Sturm R, Gresenz C, Sherbourne C, et al. The design of Healthcare for Communities: a study of health care delivery for alcohol, drug abuse, and mental health conditions. Inquiry 1999;36(2):221–233 [PubMed] [Google Scholar]
- 9.Orzechowski W, Walker RC. The Tax Burden on Tobacco Vol 42 Arlington, VA: Orzechowski and Walker; 2007 [Google Scholar]
- 10.Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition Washington, DC: American Psychiatric Association; 1987 [Google Scholar]
- 11.Saunders JB, Aasland OG, Babor TF, Fuente JR, Grant M. Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection of Persons With Harmful Alcohol Consumption-II. Addiction 1993;88(6):791–804 [DOI] [PubMed] [Google Scholar]
- 12.Health Services Research Center, University of California, Los Angeles Healthcare for Communities. Household survey public use files: user's guide & codebook (round two, release 1). 2004. Available at: http://www.hsrcenter.ucla.edu/research/hcc/userguide.pdf. Accessed May 24, 2009