Figure 1. Cumulative prevalence of alcohol use stages in the total sample (N = 10 123).
Figure 2. Cumulative prevalence of drug use stages in the total sample (N = 10 123).
Customize your JAMA Network experience by selecting one or more topics from the list below.
Swendsen J, Burstein M, Case B, et al. Use and Abuse of Alcohol and Illicit Drugs in US AdolescentsResults of the National Comorbidity Survey–Adolescent Supplement. Arch Gen Psychiatry. 2012;69(4):390–398. doi:10.1001/archgenpsychiatry.2011.1503
Context Comprehensive descriptions of substance use and abuse trajectories have been lacking in nationally representative samples of adolescents.
Objective To examine the prevalence, age at onset, and sociodemographic correlates of alcohol and illicit drug use and abuse among US adolescents.
Design Cross-sectional survey of adolescents using a modified version of the Composite International Diagnostic Interview.
Setting Combined household and school adolescent samples.
Participants Nationally representative sample of 10 123 adolescents aged 13 to 18 years.
Main Outcome Measures Lifetime estimates of alcohol and illicit substance use and DSM-IV diagnoses of abuse, with or without dependence.
Results By late adolescence, 78.2% of US adolescents had consumed alcohol, 47.1% had reached regular drinking levels defined by at least 12 drinks within a given year, and 15.1% met criteria for lifetime abuse. The opportunity to use illicit drugs was reported by 81.4% of the oldest adolescents, drug use by 42.5%, and drug abuse by 16.4%. The median age at onset was 14 years for alcohol abuse with or without dependence, 14 years for drug abuse with dependence, and 15 years for drug abuse without dependence. The associations observed by age, sex, and race/ethnicity often varied significantly by previous stage of use.
Conclusions Alcohol and drug use is common in US adolescents, and the findings of this study indicate that most cases of abuse have their initial onset in this important period of development. Prevention and treatment efforts would benefit from careful attention to the correlates and risk factors that are specific to the stage of substance use in adolescents.
The patterns of alcohol and drug use that emerge during adolescence are increasingly recognized as important determinants of later substance use behavior and associated disorders.1-9 Estimates from major surveys in the United States, including the National Survey on Drug Use and Health (NSDUH),10 the National Health and Nutrition Examination Surveys (NHANES),11 and the Monitoring the Future study,12 indicate that by 17 years of age most adolescents (59% to 71%) had consumed alcohol, 31% to 44% had tried cannabis, and 4% to 6% had tried cocaine. Although each of these investigations provides information on the prevalence of alcohol or drug use among adolescents, only the NSDUH examined substance use disorders, and the assessments were limited to symptoms experienced during the previous 12 months. The only national survey to use direct diagnostic interviews regarding both current and lifetime substance use among US adolescents recently described large increases in the cumulative incidence of substance use disorders between 13 and 18 years of age13 and an overall lifetime prevalence that approaches that of adult samples.14-17 Nationally representative data concerning the full trajectory of substance use and associated disorders in adolescence have been lacking and are necessary for attaining further progress in prevention and public health objectives.18
In addition to more complete descriptions of substance use trajectories in adolescents, assessment of factors associated with the development of substance use disorders in nationally representative samples is crucial. Epidemiologic investigations of adults have consistently demonstrated a higher prevalence of abuse and dependence among males,14-20 as well as decreased prevalence of these disorders among certain minority ethnic subgroups.14-17,21 Despite the marked increase in the prevalence of substance use disorders from early to late adolescence,13 it is uncertain to what degree these sociodemographic differences may characterize recent adolescent cohorts. For example, no salient sex differences were observed for initial alcohol or illicit drug use in NHANES between 1999 and 2004, and ethnic differences for illicit substances in this investigation were at times contrary to the adult literature.11 An additional rationale for understanding sociodemographic variation is the need for information concerning the timing of their influence across the substance use trajectory. Recent analyses of population-based surveys have shown that the magnitude and even direction of associations observed for many sociodemographic variables are specific to stages of use.16,17,22,23 Analyses of sociodemographic factors conditioned on stage of previous use are currently lacking in US adolescent samples, and little is known about the timing of their expression during this key stage of development.
The present investigation examines these issues in the National Comorbidity Survey–Adolescent Supplement (NCS-A), a nationally representative survey that used direct diagnostic interviews with youth aged 13 to 18 years. The goals of this report are to (1) describe the prevalence of 4 lifetime stages of alcohol use (use, regular use, abuse without dependence, and abuse with dependence) and 4 lifetime stages of illicit drug use (opportunity to use, first drug use, drug abuse without dependence, and drug abuse with dependence); (2) estimate the prevalence of these stages by age, sex, and race/ethnicity; and (3) examine the association of these sociodemographic variables with the risk of transition among stages.
The NCS-A is a nationally representative, face-to-face survey of 10 123 adolescents aged 13 to 18 years in the continental United States conducted between February 1, 2001, and January 30, 2004.24 The survey was administered by the professional interview staff of the Institute for Social Research at the University of Michigan, Ann Arbor. There were 197 interviewers supervised by a team of 18 experienced regional supervisors. A study manager located at the central facility in Michigan oversaw the work of the supervisors and their staff. The NCS-A was performed in a dual-frame sample that included a household subsample and a school subsample.24-27 The overall NCS-A response rate combining the 2 subsamples was 82.9%. Comparisons of sample and population distributions on US Census sociodemographic variables, and in the school sample on school characteristics, documented only minor differences, which were corrected with poststratification weighting. The recruitment and consent procedures were approved by the Human Subjects Committees of both Harvard Medical School and the University of Michigan. On making in-person contact, the interviewer answered questions before obtaining written informed consent from the parent and then written informed assent from the adolescent. Once the survey was completed, cases were weighted for variation in within-household probability of selection (in the household subsample) and for residual discrepancies between the sample and the US population on the basis of sociodemographic and geographic variables. These weighting procedures are discussed in more detail elsewhere.26,27 Sociodemographic variables examined in this report include age (in years), sex, and self-described race/ethnicity. Approximately half the sample was male (51.3%) and the mean age was 15.2 years, with a larger proportion of adolescents aged 13 to 14 years (36.2%) and approximately equal distributions of adolescents aged 15 to 16 and 17 to 18 years. The sample was composed of 65.6% non-Hispanic whites, 15.1% non-Hispanic blacks, and 14.4% Hispanics.
Adolescents were administered a modified version (version 3.0) of the World Health Organization Composite International Diagnostic Interview (CIDI), a fully structured interview administered by trained lay interviewers to generate DSM-IV diagnoses.28 The modifications to the CIDI were conducted following a standard 4-step process of instrument development proposed by experts in survey methods.24 First, diagnostic sections of the adult CIDI were reviewed to determine whether assessment for particular disorders should be removed based on low presumed prevalence in adolescents. Second, the language in the remaining CIDI sections was modified to enhance comprehension with adolescents using an iterative process. Third, CIDI modules were modified in content to make them more germane to the contexts and experiences of adolescents. The most common change of this type required altering references from adult contexts (eg, work life and parenting) to adolescent contexts (eg, school life and peer relationships). Fourth, the finalized revision of each diagnostic module was reviewed for meaning, logic, and comparability to the adult version. Each diagnostic section was then systematically piloted to test the flow and timing among adolescents, with subsequent modifications to reduce the length of the diagnostic sections. Lifetime substance use disorders assessed in the CIDI included alcohol and drug abuse, and individuals fulfilling abuse criteria were administered questions concerning dependence criteria. This report therefore presents both categories of abuse (with or without dependence). In light of differences in the pertinence of diagnostic categories for nicotine, trajectories for this substance will be addressed in a separate manuscript. In addition to DSM-IV diagnoses, all respondents were asked about the use of diverse substances and their age at which specific stages of use first occurred. For alcohol, the lifetime use question concerned age at which respondents first (if ever) had a drink with alcohol, specified as beer, wine, wine coolers, and hard liquor (eg, vodka, gin, whiskey, and mixed drinks). First regular use of alcohol was defined as the age at which the respondent first had at least 12 drinks within a single year. For illicit drugs, participants were asked their age at which they had first (if ever) consumed a range of specific substances, including marijuana or hashish; cocaine in any form (eg, powder, crack, freebase, coca leaves, or paste); tranquilizers, stimulants, pain killers, or other prescription drugs either without the recommendation of a health professional or for any reason other than a health professional said they should be used; and heroin, opium, glue, lysergic acid diethylamide, peyote, or any other drug. Participants were also asked about the first time they had an opportunity to drink alcohol or use drugs, regardless of whether they used them. Opportunity to use was defined as when someone either offered them alcohol or drugs or when the individual was present when others were using and could have used if he or she wanted to. The age at first opportunity to use substances was recorded separately for alcohol and drugs. In light of the high availability of legal substances, such as alcohol, only opportunity to use illicit drugs is examined in this investigation.
Cross-tabulations were used to calculate the prevalence of the 4 stages of alcohol and drug use by demographic characteristics. Conditional prevalence estimates for each alcohol or drug use stage were also calculated among those who had reached the earlier stage of use. Estimated projections of the cumulative probability of stages of alcohol or drug use as of the age of 18 years were obtained by the actuarial method implemented in PROC LIFETEST in SAS statistical software (version 9.2; SAS Institute, Inc). The associations of age, sex, and race/ethnicity with the stages of alcohol or drug use were examined using multiple logistic regression analysis. Because the NCS-A data are both clustered and weighted, the Taylor series linearization method implemented in SUDAAN (version 10; Research Triangle Institute) was used to estimate standard errors of logistic regression coefficients. The coefficients and standard errors were exponentiated to produce odds ratios and 95% CIs. Significance of predictor sets was evaluated using Wald F tests based on design-adjusted coefficient variance-covariance matrices. Statistical significance was consistently evaluated using 2-sided tests with a .05 significance level.
Alcohol use, regular use, and abuse were highly prevalent among US adolescents (Table 1). Most of the sample (59.8%) reported alcohol use at some point in their lifetime, ranging from 42.5% for adolescents aged 13 to 14 years to 78.2% for adolescents aged 17 to 18 years. Nearly 1 in 10 adolescents reported regular alcohol use in the 13- to 14-year age cohort, a figure that increased to nearly half (47.1%) of adolescents aged 17 to 18 years. Concerning alcohol use disorders, 5.2% of all US adolescents met DSM-IV criteria for alcohol abuse without dependence and 1.3% for alcohol abuse with dependence. The prevalence of alcohol abuse with or without dependence ranged from 1.3% for the youngest cohort to 15.1% for adolescents 17 to 18 years old. The cumulative prevalence of alcohol use stages by age is presented in Figure 1. The median age at onset of each category was 13 years for first alcohol use and 14 years for regular use or abuse (with or without dependence).
Male and female adolescents had essentially equivalent rates of any alcohol use, but male adolescents demonstrated slightly higher rates of regular use and abuse. These sex differences were larger in individuals aged 17 to 18 years than in the 2 younger cohorts. Black and other racial/ethnic groups generally reported lower rates of alcohol use, regular use, and abuse than either Hispanic or white adolescents. These racial differences were observed for each age cohort.
The conditional prevalence rates presented in Table 1 demonstrate salient effects for age in transitions among the stages of alcohol use. Only a few (22.8%) of the youngest users progressed to regular alcohol use, and 13.2% of this latter group developed abuse with or without dependence. By comparison, most adolescents (60.2%) 17 to 18 years old who used alcohol later reported regular alcohol use, and nearly 1 in 3 regular users (32.1%) in this age cohort met criteria for lifetime alcohol abuse with or without dependence. Regarding sex differences, male adolescents were more likely to make transitions among the stages, especially among the oldest adolescents. In addition, generally robust racial/ethnic differences were observed in the proportion of individuals making transitions from one alcohol stage to another. The percentage of individuals at each stage making transitions to the next was generally lowest for the black and other ethnic groups and was observable across all age cohorts.
Most US adolescents (60.2%) reported having had the opportunity to use illicit drugs (Table 2). This exposure concerned 39.2% of adolescents aged 13 to 14 years and 81.4% of adolescents aged 17 to 18 years, thereby doubling during the adolescent period. The use of illicit drugs for these same cohorts ranged from 9.6% (nearly a quarter of exposed youth aged 13-14 years) to 42.5% of the oldest adolescents (52.2% of those exposed), respectively. Abuse of illicit drugs was slightly more common than the abuse of alcohol, ranging from 3.4% of youth in the 13- to 14-year cohort to 16.4% of adolescents aged 17 to 18 years old. Figure 2 presents the cumulative prevalence of drug use stages by age. The median ages of onset for each category were 13 years for the first opportunity to use illicit drugs, 14 years for the first illicit drug use, 14 years for drug abuse with dependence, and 15 years for drug abuse without dependence. Similar to the pattern in alcohol use, there was a slight overall male preponderance in prevalence of stages of drug use, which was most pronounced among older adolescents. The same pattern of racial/ethnic differences was observed across the 3 age cohorts, with black and other racial/ethnic groups having typically lower rates than white or Hispanic adolescents. High rates of opportunity to use drugs and drug use were observed among Hispanic adolescents, especially in the youngest age cohort. Concerning conditional prevalence rates, 40.5% of the sample used illicit drugs when given the opportunity, and more than one-third of adolescents who used illicit drugs developed abuse (36.6%). The overall patterns of effects by sex and race/ethnicity in the conditional prevalence rates for drug use stages were highly similar to those observed for alcohol, as demonstrated by increasing male preponderance in transitions among older adolescents and generally lower rates of transitions, particularly among black adolescents. Concerning types of illicit drugs, cannabis was the most frequently used substance in all age cohorts (Table 3). The use of each category of illicit substance was again typically lowest for black adolescents and slightly higher for male adolescents compared with female adolescents, with the exception of prescription drug use.
Adolescent age cohort, sex, and race/ethnicity in the risk of transition among the alcohol and drug use stages were examined in multivariate analyses (Table 4). For alcohol, the 2 younger age cohorts were less likely than adolescents aged 17 to 18 years to become regular drinkers if they had already consumed this substance and less likely to develop abuse (without dependence) if they had reached regular use levels. Female adolescents were also less likely to make the transition from any alcohol use to regular use, as were adolescents who were black or members of other racial/ethnic groups compared with white adolescents. However, the patterns observed for drug use categories were more complex in the multivariate analyses than those observed for alcohol. The 2 younger cohorts were less likely than older adolescents to use drugs when given the opportunity, but no significant sex differences were observed in the risk of transition among drug use stages. Black adolescents were less likely than white adolescents to use drugs when given the opportunity and less likely to abuse drugs (without dependence) if they were drug users. Conversely, adolescents of other racial/ethnic groups were more likely than whites to develop abuse with dependence if they were already drug users.
Knowledge of the developmental trajectories of alcohol and illicit drug use among US adolescents has relied on retrospective reports from nationally representative adult samples and on prospective observation of clinical cohorts and local community youth samples. The purposes of the present investigation were to estimate in a nationally representative sample of adolescents the prevalence of multiple states of alcohol and illicit drug use and abuse and to examine the association of sociodemographic variables with transitions across these substance use categories.
The present study reports several novel findings concerning the incidence and prevalence of alcohol and drug abuse among US youth. Although the NSDUH has shown that the rates of abuse with or without dependence of alcohol or illicit drugs have converged during the past several years among adolescents aged 12 to 17 years, we found greater rates for drug abuse (8.9%) than alcohol abuse (6.5%) among adolescents. Our conditional analyses also revealed high rates of regular use and abuse in adolescent substance users. Among alcohol users, 42.2% reported regular use, and 25.9% of these regular drinkers met criteria for abuse with or without dependence. For illicit substances, 40.5% reported using drugs after having had the opportunity to do so, and nearly 36.6% of drug users met criteria for abuse with or without dependence. These conditional rates are higher than base (unconditional) rates reported by prior surveys of youth and therefore provide additional information concerning risk of substance abuse after initial use during adolescence. The risk of drug abuse among adolescent users is of concern considering the recent findings from the Monitoring the Future and NSDUH, which show a resurgence of marijuana use among adolescents, even surpassing use of nicotine, a finding that may reflect increasingly tolerant attitudes concerning the use of illicit drugs.10,12
The median age at onset of alcohol abuse with or without dependence was 14 years, with 15.1% meeting diagnostic criteria by 18 years of age. Drug abuse was characterized by a similarly young median age at onset and with a prevalence of 16.4% among the oldest adolescents. These adolescent lifetime prevalence estimates for abuse can be compared with adult prevalences observed in the National Comorbidity Survey–Replication, a study that used parallel methods and was conducted concurrently with the NCS-A. Lifetime prevalence of alcohol abuse in the adult survey (18.6%) was only marginally higher than the adolescent prevalence we observed, whereas the prevalence of drug abuse among adults (11.1% overall and 15.1% for the youngest cohort aged 19-30) was actually lower than the adolescent prevalences we estimated.16,17
Taken together, these findings suggest greater incidence risk and lifetime prevalence of substance abuse during adolescence than has been estimated from adult samples.14-17 Such differences are likely to reflect the inclusion of older-onset cases in investigations of adults and increases in retrospective memory biases in these samples. It also remains possible that younger cohorts are more likely to use specific substances or to use them at an earlier age. In this regard, the conditional analyses demonstrate that the risk of abuse among users is particularly high among the youngest adolescents. The conditional rate for abuse with or without dependence among the youngest drug users (13-14 years old) is 10-fold the unconditional rate (35.4% vs 3.4%). By contrast, the conditional rate of abuse with or without dependence among the oldest adolescents is just more than twice the unconditional rate (38.6% vs 16.4%). Although high conditional rates in younger adolescents may be partly attributable to the low unconditional prevalences, they nonetheless pinpoint early adolescence as a period of heightened risk for substance use and abuse.29,30 The elevated rate and risk of substance use and abuse among youth should be considered relative to neuroimaging data that demonstrate the occurrence of important maturational changes in the brain during adolescence, particularly in regions associated with regulatory control and decision making. Such normative developmental processes may be substantially compromised by substance use.31-35
Although probability of each stage of alcohol and drug use increased with age, rates were almost always lowest for black and other racial/ethnic groups compared with white or Hispanic adolescents. These findings may contradict common assumptions but are consistent with previous investigations in adult and adolescent samples.10,11 No salient sex differences were observed in lifetime alcohol use, whereas minor differences were observed for all other stages of alcohol or drug use among adolescents aged 13 to 16 years. However, for those 17 to 18 years of age, male adolescents had notably higher rates for most stages of alcohol or drug use. These findings are globally consistent with previous reports of the associations between sociodemographic characteristics and substance use and substance use disorders11,12,14,16-21,29 and indicate that sex differences in substance use behavior are most visible after 15 years of age. In conditional analyses, male alcohol users were significantly more likely to transition to regular alcohol use, whereas no sex differences were associated with drug use transitions. Reduced risk for black adolescents was significant only in transitions from alcohol use to regular use, from having the opportunity to use drugs to drug use, and from drug use to drug abuse without dependence. Despite lower rates of drug use among other racial/ethnic groups, these adolescents were more likely than white adolescents to transition from use to abuse with dependence. These findings are concordant with previous studies16,17,22,29 demonstrating that stage of previous substance use moderates the role of sociodemographic factors.
The principal strengths of this study include its use of a nationally representative adolescent sample with direct diagnostic interviews and the assessment of multiple stages of alcohol and drug use. Study limitations include its cross-sectional design, retrospective reporting for age at onset information, and the lack of assessment of alcohol or drug dependence without a history of abuse. However, although we would expect that only a small proportion of youth would be excluded in this latter reason, the conclusions are limited to the categories examined and may not be generalizable outside of the US adolescent population. Caution is also warranted in drawing parallels to findings from retrospective estimates in adult samples because such interviews may underestimate rates due to forgetting or other memory biases and may increase biases in estimates of age of disorder onset due to the longer period of recall.36
In addition, the use of the term lifetime to describe the prevalence of substance use disorders in adolescence does not necessarily indicate the persistence of these disorders in the long term, in particular because individuals may “mature out” of harmful substance use as they progress into adulthood or begin to assume adult responsibilities.37,38 A final limitation of the present investigation is that it cannot account for societal factors that vary over time and that affect prevalence rates, such as the degree of tolerance for specific illicit substances or their availability. It is important to consider that psychoactive substances are implicated in more than 12% of mortality worldwide and their use constitutes the leading cause of preventable death.39,40 Because the early onset of substance use is a significant predictor of substance use behavior and disorders in a lifespan,1-9 the public health implications of the current findings are far reaching. The prevention of both alcohol and illicit drug abuse requires strategies that target early adolescence and take into account the highly differential influence that population-based factors may exert by stage of substance use. The lack of information concerning the precise stage of alcohol or drug use most associated with sociodemographic factors may partly explain the modest success observed for certain prevention strategies.41-43 An additional challenge for national policy concerns the emergence of drug instrumentalization models that emphasize the adaptive role that psychoactive substances may play in the daily lives of nondependent individuals.44 The critical evaluation of such perspectives and the development of appropriate national policy on alcohol and drug use can only be achieved through empirical data describing the nature of harm or risk posed by alcohol and drugs.
Correspondence: Kathleen R. Merikangas, PhD, Genetic Epidemiology Research Branch, Intramural Research Program, National Institute of Mental Health, 35 Convent Dr, MSC 3720, Bethesda, MD 20892-3720 (firstname.lastname@example.org).
Submitted for Publication: June 7, 2011; final revision received August 11, 2011; accepted September 4, 2011.
Financial Disclosure: None reported.
Funding/Support: This work was supported by grant Z01 MH002808-08 from the Intramural Research Program of the National Institute of Mental Health. The NCS-A and the larger program of related National Comorbidity Surveys are supported by grant U01-MH60220 from the National Institute of Mental Health.
Disclaimer: The views and opinions expressed in this article are those of the authors and should not be construed to represent the views of any of the sponsoring organizations, agencies, or US government.
Additional Information: Jianping He, MSc, performed all statistical analyses for this investigation.