TOBACCO USE IS INCREASING

Transcription

1 ORIGINAL CONTRIBUTION US College Students Use of Tobacco roducts Results of a National Survey Nancy A. Rigotti, MD Jae Eun Lee, DrH Henry Wechsler, hd TOBACCO USE IS INCREASING among young Americans. Cigarette smoking rates among adolescents increased by 32% between 1991 and ,2 Cigarette smoking by young adults (18-24 years) increased by 16% between 1995 and 1997, presumably reflecting the aging of the adolescent cohort. 2 Young adults, who represent the youngest legal targets of tobacco industry marketing, may also be initiating tobacco use in larger numbers. If this trend continues, it threatens to reverse the decline in smoking prevalence among US adults that has occurred during the past half century. 3 The smoking behavior of college students is a useful index of tobacco use among young adults. More than one third of adults aged 18 to 24 years attend college, and one quarter attend a 4-year college. 4 US colleges and universities enroll more than 12 million students 5 and are a potential site for interventions to discourage tobacco use. Between 1993 and 1997, cigarette smoking prevalence increased 28% among US college students. 6 Whether this trend is continuing needs to be determined. There is also reason to suspect that college students use a broader range of tobacco products than cigarettes, although this has not been assessed. 7 Cigar smoking is a particular concern. 8 Following a sharp increase in promotional Context Adults aged 18 to 24 years, many of whom are in college, represent the youngest legal targets for tobacco industry marketing. Cigarette smoking has been described among college students, but little is known about noncigarette tobacco use by college students or cigar use by adults of any age. Objectives To assess the prevalence of all forms of tobacco use (cigarettes, cigars, pipes, and smokeless tobacco) among US college students and to identify studentand college-level factors associated with use of each product. Design The Harvard College Alcohol Survey, a self-administered survey conducted in Setting One hundred nineteen nationally representative US 4-year colleges. Subjects A total of randomly selected students (60% response rate). Main Outcome Measures Self-report of current (in the past 30 days), past-year, and lifetime use of cigarettes, cigars, pipes, smokeless tobacco, and all tobacco products. Results Nearly half (45.7%) of respondents had used a tobacco product in the past year and one third (32.9%) currently used tobacco. Cigarettes accounted for most of the tobacco use (28.5% current prevalence), but cigar use was also substantial (37.1% lifetime prevalence, 23.0% past-year prevalence, and 8.5% current prevalence) and combinations of the 2 were common. Total tobacco use was higher in men than in women (37.9% vs 29.7%;.001), despite nearly identical current cigarette smoking rates between the sexes (28.5% for women vs 28.4% for men), because of greater use of cigars (current prevalence, 15.7% vs 3.9%;.001) and smokeless tobacco (current prevalence, 8.7% vs 0.4%;.001) by men. Tobacco use was significantly higher among white students (.001), users of other substances (alcohol and marijuana) (.001), and students whose priorities were social rather than educational or athletic (.05). Among students who had used both cigars and cigarettes, only 8.9% smoked cigars at an earlier age than they had smoked cigarettes. Conclusion Our study indicates that tobacco use is common among college students and is not limited to cigarettes. College appears to be a time when many students are trying a range of tobacco products and are in danger of developing lifelong nicotine dependence. National efforts to monitor and reduce tobacco use of all types should expand to focus on college students and other young adults. JAMA. 2000;284: activities by manufacturers, cigar consumption in the United States increased by 50% between 1993 and 1998, reversing a 30-year decline. 9,10 The rise Author Affiliations and Financial Disclosure are listed at the end of this article. Corresponding Author and Reprints: Nancy A. Rigotti, MD, Division of General Medicine, Massachusetts General Hospital, 50 Staniford St, 9th Floor, Boston, MA ( nrigotti@partners.org) American Medical Association. All rights reserved. (Reprinted) JAMA, August 9, 2000 Vol 284, No

2 in cigar sales presumably corresponds to an increase in cigar smoking prevalence, but this is uncertain because recent data on the prevalence of cigar use are limited and not based on national samples. 11,12 Data from California suggest that cigar use is increasing most rapidly among young adults, who had the lowest rates of cigar use before ,11 Cigar smoking rates are higher among individuals with more education and higher incomes, a pattern opposite to that observed for cigarette smoking. 2,11 High rates of cigar use were reported by adolescents in high school surveys conducted between 1996 and ,13-15 These adolescents may continue to smoke cigars as they enter young adulthood. These lines of evidence suggest that the prevalence of cigar smoking may be especially high in college students. As young adults in the process of furthering their education, college students appear to fit the demographic profile of new cigar smokers. Consequently, total tobacco use by young adults may be considerably higher than has been detected by previous analyses limited to cigarettes. 6 Little is known about the full spectrum of tobacco use among the college population or about risk factors for the use of different tobacco products. 8 This information is essential to guide the development of interventions. This article reports the prevalence and patterns of use of all tobacco products among respondents to the 1999 Harvard College Alcohol Survey, a large nationally representative sample of US college students. METHODS Sample of Colleges The Harvard College Alcohol Study surveyed randomly selected, crosssectional samples of students in a nationally representative sample of 4-year, US colleges in 1993, 1997, and Details of the method and sampling procedures are published elsewhere. 16,17 In 1999, 128 (91%) of the original 140 colleges that participated in the 1993 study were resurveyed. articipating colleges were sent guidelines for drawing a random sample of undergraduates from their total enrollment of fulltime students. A sample of 225 was selected from each college. Nine (7%) of the 128 schools were dropped from analyses because of low response rates (less than 40% of eligible students responding). The final sample was 119 colleges. Cigarette and cigar use rates did not differ significantly between these 119 colleges and the total sample of 128 colleges, indicating that omission of the 9 schools had no effect on the major variable under study. The 119 colleges are located in 39 states (24% in the Northeast, 29% in the South, 29% in the North Central, and 18% in the West). Seventy-one percent are located in urban or suburban settings, and 29% are in small town or rural settings. More than two thirds are public institutions. Forty-four percent have more than students, 23% have 5001 to students, and 34% have less than 5000 students. Fifteen percent have a religious affiliation. Five percent enroll only women. Sample of Students Questionnaires were mailed to students at the 119 schools in February Three separate mailings were sent: a questionnaire, a reminder postcard, and a second questionnaire. Responses were anonymous. Cash awards were used to encourage students to respond. Sixty percent of students (n=14138) returned questionnaires. Response rates varied between 40% and 83% among the 119 colleges. To explore the potential for bias due to student nonresponse, a short form of the questionnaire that included a smoking question was mailed to a sample of students who did not return the 1999 questionnaire. ast 30-day cigarette smoking rates did not differ significantly between respondents to the short survey and respondents to the entire student survey (25% vs 28%, =.17). The correlations between individual colleges response rates and their 30- day cigar and cigarette prevalence were 0.04 (=.68) and 0.01 (=.11), respectively, indicating little chance for bias on this basis. Sixty-one percent of the respondents were female compared with 53% of the full-time undergraduate population of the 119 participating schools. 18 Respondents ethnicity (76% non- Hispanic white, 7% Hispanic, 6% black, and 8% Asian) closely resembles that of all students at the 119 schools (76% white, 7% Hispanic, and 5% black). 18 Eighty-five percent of respondents were 18- to 24-year-olds; 23% were freshmen, 22% sophomores, 25% juniors, 22% seniors, and 8% fifth-year or postgraduate students. Respondents age distribution closely matches that of students at all 4-year colleges. 19 Questionnaire The questionnaire assessed demographic and background characteristics; tobacco, alcohol, and other substance use; satisfaction with education; and students interests and lifestyle choices. 16,17 To assess tobacco use, respondents were asked if they had smoked a cigarette, cigar, or pipe or used smokeless tobacco. For each, response options were never used, used, but not in the past 12 months, used, but not in the past 30 days, or used in the past 30 days. Also assessed were the number of days in the past 30 days in which a student had smoked cigarettes and cigars and the age at which they first smoked these products. Data Analysis For each tobacco product, all students who reported any use were considered to be ever (lifetime) users. Current users were students who used the product in the past 30 days. ast-year users reported having used the product in the past year. Total tobacco use included cigarettes, cigars, smokeless tobacco, and pipes. Binge drinking was defined as drinking 5 or more drinks in a row for men and 4 or more drinks in a row for women during the previous 2 weeks. 20 Current marijuana users were students who had smoked marijuana in the past 30 days. Other illicit drug use was not included in the analysis, because current use rates were very low. 700 JAMA, August 9, 2000 Vol 284, No. 6 (Reprinted) 2000 American Medical Association. All rights reserved.

3 Statistical analyses were carried out using SAS statistical software (SAS Institute Inc, Cary, NC). Univariate analyses identified student-level and collegelevel characteristics associated with current cigar, cigarette, smokeless tobacco, and total tobacco use. Sociodemographic factors included age, sex, ethnicity, year in school, marital status, and highest parental educational attainment (a proxy for socioeconomic status). Other student-level variables included current marijuana use, binge drinking, grade point average, a single-item measure of satisfaction with education, an item assessing the importance to a student of participating in 9 activities (eg, athletics, academic work, fraternity or sorority life, parties, and religion), and an item measuring the number of hours per week a student spent on 8 activities (eg, playing intercollegiate sports, studying, socializing with friends). Collegelevel factors included geographic location (West, South, North Central, and Northeast), campus location (rural or small town vs urban or suburban), size of enrollment ( 5000, , or 10000), public vs private status, religious affiliation, coeducational vs allfemale, competitiveness for admission (based on percentage of applicants accepted), commuter vs noncommuter school, and survey response rate. Student-level and college-level factors that were significantly associated with cigar, cigarette, smokeless tobacco, and total tobacco use in the univariate analysis (.001) were included in multiple logistic regression models. Separate models were built for cigar, cigarette, smokeless tobacco, and total tobacco use. Current use was the outcome variable for all models. All multivariate models included sex, ethnicity, year in school, marital status, and highest parental educational attainment. Two-way interaction terms of these factors with sex were included in the initial model, and those with statistical significance were retained in the final model. Year in school was used in place of age to examine the effect of duration of exposure to the college environment. Results do not differ when age Table 1. revalence of of All Types All, % (N = ) is substituted in the models. Adjusted odds ratios (ORs) with 95% confidence intervals are reported. The final logistic regression model was fitted by the Generalized Estimating Equations approach, 21,22 with the independent working correlation, to make more robust inference using clustered outcomes due to our sampling scheme. Odds ratios obtained from the Generalized Estimating Equations method were almost identical with those from ordinary logistic regression, whereas the SEs of ORs associated with the college characteristics were slightly larger. The Generalized Estimating Equations based results are reported herein. Men, % (n = 5471) Women, % (n = 8667) * Total tobacco Current use ast year use Lifetime use Cigarettes Current use ast year use Lifetime use Cigars Current use ast year use Lifetime use Smokeless tobacco Current use ast year use Lifetime use ipes Current use ast year use Lifetime use * values derived by 2 test for trend to compare men vs women on each item. RESULTS revalence of More than 60% of college students have tried a tobacco product, nearly half used tobacco in the past year, and one third used tobacco in the past month (TABLE 1). Cigarettes account for most of the tobacco use. More than half of college students (53.4%) have smoked a cigarette, 38.1% did so in the past year, and 28.5% were current (past 30-day) cigarette smokers. Among current smokers, 32.0% smoke less than 1 cigarette per day, 43.6% smoke 1 to 10 cigarettes per day, and only 12.8% smoke 1 or more pack per day. Current, pastyear, and lifetime cigarette smoking rates in this college sample did not change between 1997 and Total tobacco use is higher among men than women (Table 1). More than half of men (53.0%) and 41.3% of women have used tobacco in the past year, whereas 37.9% of men and 29.7% of women are current tobacco users. The sex difference in total tobacco use is entirely attributable to a higher prevalence of noncigarette tobacco use among men, because men and women have nearly identical cigarette smoking rates. Cigar smoking accounts for the largest share of noncigarette tobacco use (Table 1). More than one third of college students have ever smoked a cigar, including more than half of the men and one quarter of the women. Nearly 1 in 4 college students smoked a cigar in the past year, and 8.5% smoked a cigar in the past month. Most cigar use is occasional: less than 1% of current (past 30-day) cigar users smoked daily, and 89.7% smoked a cigar on fewer than 5 of the past 30 days. Smokeless tobacco use is much less prevalent than cigarette or cigar smoking. ipe smoking is rare American Medical Association. All rights reserved. (Reprinted) JAMA, August 9, 2000 Vol 284, No

4 Table 2. revalence of Current by Race and Sex* White, % (n = ) Hispanic, % (n = 1018) The use of cigarettes, smokeless tobacco, and all products combined is more common in non-hispanic whites than in Hispanics, blacks, or Asians (TABLE 2). In contrast, the prevalence of cigar use is similar in whites and blacks (9.2% vs 8.1%) but lower in Hispanics and Asians. Cigar smoking is actually more prevalent among black women than white women (6.8% vs 4.0%, =.003), whereas a reverse relationship exists for men (10.9% vs 17.2%, =.008). This finding should be interpreted with caution because results for blacks are based on a small sample and need confirmation. Combinations of Tobacco roducts Most college students who have used tobacco used more than 1 product. Among tobacco users, 51.3% used more than 1 tobacco product in the past year (36.3% used 2 tobacco products, 14.4% used 3 products, and 0.6% used all 4 products). The most frequent combinations were cigarettes and cigars (19.7%), cigarettes and pipes (12.0%), and cigarettes, cigars, and smokeless tobacco (6%). In contrast, 77.5% of current tobacco use was restricted to a Asian, % (n = 1117) Black, % (n = 788) Total tobacco All Men Women Cigarettes All Men Women Cigars All Men Women Smokeless tobacco All Men Women ipes All Men Women *Referent is non-hispanic white. values derived by 2 test. single product. Among current users, the most frequent combination was cigars and cigarettes. Men used multiple tobacco products more often than women; 58.2% of men vs 27.5% of women had used more than 1 tobacco product in the past year (.001), whereas 37.1% of men but only 10.6% of women had used more than 1 tobacco product in the past month (.001). The combination of cigarettes and cigars accounts for nearly all multiple tobacco product use in women and is by far the most frequent combination among men. Noncigarette tobacco use occurs more often in combination with cigarette use than in isolation. Among current cigar smokers, 33.4% smoked only cigars, whereas 61.4% smoked both cigars and cigarettes. Of current smokeless tobacco users, 30.6% used only smokeless tobacco, whereas 62.3% also smoked cigarettes. Age of First The median age of first cigarette use was 14 years for both sexes. The median age of first cigar use was 17 years for men and 18 years for women. However, 26.8% of cigar smokers (33.3% of women and 21.9% of men) had their first cigar at 19 years or older, when they were presumably in college. Among 5262 respondents who had ever used both cigars and cigarettes, 63.8% had smoked a cigarette at a younger age than they had smoked a cigar, 27.3% had started both at the same age, and only 8.9% had used cigars at a younger age than cigarettes. Correlates of TABLE 3 displays the multivariate analyses examining student-level and collegelevel sociodemographic factors associated with current use of all tobacco products. TABLE 4 displays the result of a similar analysis focused on student behaviors and lifestyles. Total tobacco use is significantly (.05) linked with sex, ethnicity, marital status, other substance use, high-risk behaviors, and certain college lifestyles. Male and white students are more likely to use tobacco than female and nonwhite students (Table 3). Students who use tobacco are also more likely to smoke marijuana, binge drink, have more sexual partners, have lower grades, rate parties as important, and spend more time socializing with friends (Table 4). Tobacco users are less likely than nonusers to rate athletics or religion as important and to be satisfied with their education. Total tobacco use is not related to parental educational attainment. The only collegelevel factor associated with tobacco use is geographic location. Tobacco use is lower in western colleges (Table 3). Except for sex, these same factors are also independently associated with current cigarette smoking. Cigar smokers differ from cigarette smokers in several ways. Men are far more likely than women to smoke cigars. Blacks are as likely to smoke cigars as are whites. Like cigarette smoking, current cigar use is associated with the use of other tobacco products and other substances (alcohol and marijuana) and is more common among students who express strong interest in parties. Unlike cigarettes, cigars are also 702 JAMA, August 9, 2000 Vol 284, No. 6 (Reprinted) 2000 American Medical Association. All rights reserved.

5 more often used by students who rate fraternities and sororities and attending sporting events as important. Smokeless tobacco use also has distinctive features. It has the strongest association with male sex and white race of all tobacco products. Unlike cigarettes, it is just as common in the West as in other areas, and there is a trend toward greater use by students enrolled in colleges in rural areas or small towns. Unlike cigarettes, smokeless tobacco use is more common among intercollegiate athletes. COMMENT This study contains several new findings. It demonstrates that tobacco use among college students is more prevalent than previously appreciated, because tobacco use is not limited to cigarettes. Cigar smoking is substantial, and smokeless tobacco (and, rarely, pipes) are also used. Most tobacco users use more than 1 tobacco product, with cigars and cigarettes being the most common combination. Because men are more likely than women to use noncigarette tobacco products, overall tobacco use rates are higher in men than women, despite identical rates of cigarette smoking. Whether the gap between the sexes in noncigarette tobacco product use will follow the pattern of cigarettes and narrow in the future deserves to be monitored. In the meantime, these data indicate that assessments limited to cigarette smoking will underestimate the level of tobacco use in the college population, especially among men. This study also reports some good news. Cigarette use by college students, which increased dramatically between 1993 and 1997, 6 stabilized between 1997 and Our previous work suggested college is a time of considerable flux in cigarette smoking. 6 This article extends that work to show that college students are experimenting with the full range of tobacco products and that this is occurring across all types of colleges. College students who use tobacco share many characteristics. They are more likely to be white, single, and experimenting with other risky behaviors, such as binge drinking, using marijuana, and having more sexual partners. Tobacco use also appears to be part of a college lifestyle that values social life over educational achievement, athletic participation, or religion. A similar pattern was observed in an analysis of predictors of cigarette use in the 1993 version of this survey. 23 Although many of the same characteristics identify students who are likely to use any tobacco product, this article demonstrates that there are distinctive patterns of cigar and smokeless tobacco use that distinguish their users from cigarette smokers. Cigar smoking in the United States was primarily a behavior of older men until the early 1990s. 9 Few young adults and few women smoked cigars. Our data demonstrate how much this pattern has changed. Cigar use is now common among college students. Although it remains much more common among men than women, an unprecedented number of female col- Table 3. Sociodemographic Factors Associated With Current Use of Tobacco roducts: Multiple Logistic Regression Analysis Fitted by Generalized Estimating Equations* Factor Total Cigarette Use Cigar Use Smokeless Sex Male 1.44 ( ) ( ) ( ) ( ).001 Female Referent Referent Referent Referent Race White Referent Referent Referent Referent Hispanic 0.67 ( ) ( ) ( ) ( ).002 Black 0.35 ( ) ( ) ( ) ( ).001 Asian 0.51 ( ) ( ) ( ) ( ).001 Year in school Freshman or sophomore 0.97 ( ) ( ) ( ) ( ).08 Junior or senior Referent Referent Referent Referent Marital status Single 1.45 ( ) ( ) ( ) ( ).64 Married Referent Referent Referent Referent Geographic location North Central Referent Referent Referent Referent Northeast 0.92 ( ) ( ) ( ) ( ).04 South 0.91 ( ) ( ) ( ) ( ).28 West 0.66 ( ) ( ) ( ) ( ).51 Campus location Rural or small town 1.00 ( ) ( ) ( ) ( ).07 Urban or suburban Referent Referent Referent Referent *Controlled for college response rate, commuter school, and parents highest educational attainment. indicates adjusted odds ratio; CI, confidence interval American Medical Association. All rights reserved. (Reprinted) JAMA, August 9, 2000 Vol 284, No

6 lege students now smoke cigars. Cigar use patterns had other unique features. First, cigar use by freshmen and sophomores exceeds that of juniors and seniors. This pattern was not observed for other tobacco products and was contrary to our expectation that cigar smoking rates would increase with more years of exposure to college. It suggests that the cigar use is a new phenomenon entering the college population. College student cigar use may increase in the future, as the cohort of juniors and seniors progresses through college and is joined by adolescents in whom cigar use is already high. 1,13-15 A second unexpected feature was the higher rate of cigar smoking by black students. This is in sharp contrast to black students lower rate of all other tobacco product use. This preliminary finding needs confirmation in future work with larger samples. College students who use smokeless tobacco have their own distinctive features. They are predominantly white men who attend schools in rural areas, thereby fitting the pattern of smokeless tobacco use in the broader population. 24 Intercollegiate athletes, in whom cigarette smoking rates are low, are more likely to use smokeless tobacco, perhaps as a substitute form of nicotine. Although the prevalence of tobacco use by college students is high, the intensity of their tobacco use is low. One third of cigarette smokers and 99% of cigar smokers do not smoke every day. Even this low level of use is a cause for concern. Low levels of tobacco exposure for many years can produce substantial morbidity and mortality, as research on passive smoking shows. 25 More important, the intermittent use of cigars, smokeless tobacco, or cigarettes repeatedly exposes a user to nicotine, a drug capable of producing physical dependence. Low levels of nicotine exposure during college might evolve later into nicotine dependence and daily cigarette smoking. Our data do not suggest that cigars act as a gateway to cigarette use in college students, since only 9% of cigar and cigarette users smoked cigars first, but this issue deserves further study. Cigar or smokeless tobacco use might also contribute to the persistence of tobacco use in former cigarette smokers and thereby facilitate a return to cigarette smoking. Cigar smoking among college students may diffuse to other segments of the population, especially to children or adolescents, because of the effect of peer modeling. The visibility of tobacco products on campus, even if used intermittently, sends a dangerous message about the social acceptability of tobacco use. The broad use of tobacco among a substantial portion of 18- to Table 4. Lifestyle Factors Associated With Current Use of Tobacco roducts: Multiple Logistic Regression Analysis Fitted by Generalized Estimating Equations* Factor Total Cigarette Use Cigar Use Smokeless Other tobacco use Cigarettes ( ) ( ).001 Cigars ( ) ( ).20 Smokeless ( ) ( ) Risk behaviors Marijuana use 4.62 ( ) ( ) ( ) ( ).50 Binge drinking 3.60 ( ) ( ) ( ) ( ) Sex partners in past month Education Grade point average ( B average) 1.49 ( ) ( ) ( ) ( ) ( ) ( ) ( ).21 Satisfied with education 0.71 ( ) ( ) articipation is important Art ( ) ( ).002 Fraternities or sororities ( ) ( ) ( ).17 arties 1.31 ( ) ( ) ( ) ( ).72 Athletics 0.62 ( ) ( ) ( ) ( ).11 Attending sports ( ) ( ).19 Religion 0.86 ( ) ( ) ( ).07 Spends 2 h/d Studying 0.92 ( ) ( ) ( ) ( ).23 Socializing with friends 1.13 ( ) ( ) ( ) ( ).84 Intercollegiate sports ( ) ( ).002 *Controlled for sex, race, year in school, marital status, parents highest educational attainment, geographic location, urban vs rural location, and college response rate. indicates adjusted odds ratio; CI, confidence interval; and ellipses, not included in final model. 704 JAMA, August 9, 2000 Vol 284, No. 6 (Reprinted) 2000 American Medical Association. All rights reserved.

7 24-year-olds could portend a future increase in overall adult tobacco use. The results of this study must be viewed within the context of its limitations. One potential limitation is nonresponse bias. Although the survey response rate was 60%, the prevalence of cigarette smoking did not differ between respondents and a sample of nonrespondents, and there was no correlation between smoking rates and response rates at individual schools. We attempted to control for potential bias by including college response rate as an independent variable in multivariate analyses. Students self-report was our measure of tobacco use. Biochemical measures have established the validity of relying on self-reported smoking status in national surveys. 26,27 Underreporting of smoking status is probably even less likely in this survey, which focused on alcohol rather than tobacco. College students comprise the largest group of Americans aged 18 to 24 years. This is the youngest age group that tobacco manufacturers can legally target for marketing efforts. The college years are a crucial period in the development or abandonment of smoking behavior. 6 Tobacco use in this group should be monitored closely, and young adults should be included in all tobacco control efforts. Colleges offer a potential site for interventions to discourage tobacco use. These efforts clearly need to broaden beyond cigarettes to address the use of cigars and smokeless tobacco. Colleges alcohol and substance use prevention and treatment programs should address tobacco, because use patterns are so highly correlated. Data from this study can help colleges target interventions to the students at greatest risk of using specific tobacco products. These efforts should be accompanied by environmental and policy changes that discourage tobacco use and reinforce the message that nonsmoking is the norm. One key component is to make college buildings, including dormitories and living quarters, smoke free. This policy protects nonsmokers from passive smoke exposure and protects all students from the fire hazard of tobacco use in dormitories. It also limits the visibility and accessibility of tobacco products and may discourage initiation, help keep occasional smokers from becoming regular users, and boost the success of smokers who are trying to quit. Reducing tobacco use of all types among young adults should be a national health priority. Author Affiliations: Tobacco Research and Treatment Center, Division of General Medicine, Department of Medicine, Massachusetts General Hospital, and Harvard Medical School (Dr Rigotti), and Department of Health and Social Behavior, Harvard School of ublic Health (Drs Wechsler and Lee), Boston, Mass. Financial Disclosure: Dr Rigotti currently receives grant support from Glaxo Wellcome Inc. In the past, she received support from SmithKline Beecham Consumer Healthcare and received honoraria for lectures from McNeil, Glaxo Wellcome, and SmithKline Beecham. Funding/Support: The study was supported by a grant from the Robert Wood Johnson Foundation. Acknowledgment: The authors thank Hang Lee, hd, for biostatistical assistance. REFERENCES 1. Centers for Disease Control and revention. Tobacco use among high school students: United States, MMWR Morb Mortal Wkly Rep. 1998;47: Cigarettesmokingamongadults:UnitedStates,1997. MMWR Morb Mortal Wkly Rep. 1999;48: Tobacco use: United States, MMWR Morb Mortal Wkly Rep. 1999;48: US Bureau of the Census. Statistical Abstract of the United States, th ed. Washington, DC: US Dept of Commerce; National Center for Education Statistics. Digest of Education Statistics, Washington, DC: US Dept of Education, Office of Educational Research and Improvement; NCES publication Wechsler H, Rigotti NA, Gledhill-Hoyt J, Lee H. Increased levels of cigarette use among college students: a cause for national concern. JAMA. 1998;280: Centers for Disease Control and revention. Youth risk behavior surveillance: National College Health Risk Behavior Survey: United States, MMWR Morb Mortal Wkly Rep. 1997;46:SS Jamner MS. Cigar smoking among college students: prevalence and correlates. rev Med. 1999;29: US Department of Health and Human Services. Cigars: Health Effects and Trends. Bethesda, Md: National Cancer Institute; National Institutes of Health publication Smoking and Tobacco Control Monograph US Department of Agriculture. Tobacco Situation and Outlook Report. Washington, DC: US Dept of Agriculture, Commodity Economics Division, Economic Research Service; April Gilpin EA, ierce J. Cigar smoking in California: Am J rev Med. 1999;16: Hyland A, Cummings KM, Shopland DR, Lynn WR. revalence of cigar use in 22 North American communities: 1989 and Am J ublic Health. 1998; 88: Centers for Disease Control and revention. Tobacco use among middle and high school students: Florida, 1998 and MMWR Morb Mortal Wkly Rep. 1999;48: Centers for Disease Control and revention. Cigar smoking among high school students: United States, Massachusetts, and New York. MMWR Morb Mortal Wkly Rep. 1997;46: Centers for Disease Control and revention. Tobacco use among middle and high school students: United States, MMWR Morb Mortal Wkly Rep. 2000;49: Wechsler H, Davenport A, Dowdall G, Moeykens B, Castillo S. Health and behavioral consequences of binge drinking in college. JAMA. 1994; 272: Wechsler H, Dowdall GW, Davenport A, Castillo S. Correlates of college student binge drinking. Am J ublic Health. 1995;85: US Department of Education. Digest of Education Statistics. Washington, DC: National Center of Educational Statistics; US Bureau of the Census. Current opulation Reports, and Washington, DC: US Bureau of the Census; Wechsler H, Isaac N. Binge drinkers at Massachusetts colleges: prevalence, drinking style, time trends, and associated problems. JAMA. 1992;267: Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika. 1992;73: Zeger SL, Liang KY. Models for longitudinal data: a generalized estimating equation approach. Biometrics. 1988;44: Emmons KM, Wechsler H, Dowdall G, Abraham M. redictors of smoking among US college students. Am J ublic Health. 1998;88: Hatsukami DK, Severson HH. Oral spit tobacco: addiction, prevention and treatment. Nicotine Tobacco Res. 1999;1: US Department of Health and Human Services. The Health Consequences of Involuntary Smoking: A Report of the Surgeon General. Washington, DC: US Dept of Health and Human Services; Caraballo RS, Giovino GA, echacek TF, Mowery. Agreement between self-reports of cigarette smoking and biochemical measurement of serum nicotine levels. Abstract presented at: Annual Meeting of the Society for Research on Nicotine and Tobacco; March 28, 1998; New Orleans, La. 27. Frier ME, Bell RM, Ellickson L. Do Teens Tell the Truth? The Validity of Self-Reported by Adolescents. Santa Monica, Calif: RAND; March N-3291-CHF American Medical Association. All rights reserved. (Reprinted) JAMA, August 9, 2000 Vol 284, No