Prevalence, Correlates, Disability, and Comorbidity of DSM-IV Drug Abuse and Dependence in the United States: Title and subTitle BreakResults From the National Epidemiologic Survey on Alcohol and Related Conditions FREE

The abuse of and dependence on illicit substances are widespread among the general population and are associated with substantial societal, personal, and economic costs.^{1 - 4} National epidemiologic surveys^{5 - 8} and numerous clinical studies^{9 - 13} consistently indicate that drug use disorders have strong associations with alcohol use disorders and mood, anxiety, and personality disorders (PDs). Axis I and II comorbidity with drug use disorders has been associated with underachievement, decreased work productivity, poor health, neuropsychological impairment, human immunodeficiency virus infection, hepatitis, social dysfunction, violence, incarceration, poverty, homelessness, a lower probability of recovery, poor treatment outcome, and poor quality of life.^{14 - 18} Drug use disorder comorbidity also increases the risk of suicide attempts, especially among individuals with bipolar disorder.¹⁹

Although extensive data on drug use in the US population have been available on an ongoing basis for adults and adolescents,^{20 - 21} epidemiologic data on the prevalence, correlates, disability, treatment, and comorbidity of drug use disorders among adults are seldom collected. In fact, it has been more than 16 years since such detailed information on drug use disorders in the United States has been published. In one of those studies, the 1990-1992 National Comorbidity Survey,⁷ DSM-III-R²² criteria were used to assess drug use disorders. In another, the 1991-1992 National Longitudinal Alcohol Epidemiologic Survey,²³ DSM-IV²⁴ criteria were used, but assessments of disorders comorbid with drug use disorders were limited to major depression and dysthymia. Although the National Survey on Drug Use and Health began to collect 12-month but not lifetime prevalence data on DSM-IV drug use disorders in 2000, data on disability and specific psychiatric comorbidity were not collected.²¹

In view of the seriousness of drug use disorders and the adverse impact of their comorbidity with other psychiatric disorders, there is a pressing need for current, detailed data on the prevalence, correlates, disability, and comorbidity of drug use disorders derived from a single, uniform data source. Furthermore, the need for current information on drug use disorders using DSM-IV criteria is critical because the diagnostic definitions of drug use disorders have changed during successive revisions of the nomenclature, and these changes can influence the prevalence of disorders and their relationships with sociodemographic and clinical correlates, disability, and other psychiatric disorders. Twelve-month rates of drug abuse reported in previous epidemiologic surveys conducted worldwide since the early 1980s have remained relatively stable cross nationally, regardless of whether DSM-III,²⁵ DSM-III-R, or DSM-IV definitions were used (0.9%,⁸ 0.3%-0.8%,^{7 ,26 - 29} and 0.9%-1.1%,^{21 ,30 - 31} respectively). However, rates of 12-month drug dependence were somewhat lower when DSM-III criteria were used (1.2%),⁸ compared with studies using DSM-III-R (0.6%-2.8%)^{7 ,26 - 29} and DSM-IV (0.5%-2.0%)^{21 ,30 - 31} criteria. For lifetime drug abuse, rates were lower for DSM-III (2.6%)⁸ compared with DSM-III-R (1.5%-8.5%)^{7 ,26 - 29} and DSM-IV (0.8%-7.9%) criteria.^{21 ,30 - 31} In contrast, lifetime rates of DSM-IV drug dependence (0.4%-2.9%)^{30 - 31} were lower than those derived using DSM-III (3.5%)⁸ and DSM-III-R (0.7%-7.5%)^{7 ,26 - 29} criteria. Although these previous surveys contributed important information on drug use disorders during the 1980s and early 1990s, little is known about the epidemiology of drug use disorders since 1992.

The lack of current and comprehensive information on DSM-IV drug use disorders in the United States represents a gap in our knowledge with relevance to prevention, treatment intervention, and economic costs. Accordingly, the present study was designed to address this gap using data from the 2001-2002 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC).^{32 - 33} The NESARC assessed DSM-IV alcohol and drug use disorders, nicotine dependence, mood and anxiety disorders, and 7 of the 10 PDs. The sample size (N = 43 093) and high response rate (81.0%) of the NESARC allow for the estimation of 12-month and lifetime prevalence and comorbidity of drug abuse and dependence separately (not aggregated), especially among major sociodemographic subgroups of the population, including those for which detailed information has not been available (eg, Native Americans and Asians). Furthermore, comorbidity of drug abuse or dependence and each specific psychiatric disorder was examined while controlling for other psychiatric disorders. This information, which is necessary for understanding the unique relationship of drug abuse and dependence with other disorders while controlling for the comorbidity of these disorders with each other, has not been addressed in previous research. This study also provides information on disability and age at onset and examines the characteristics of individuals with drug abuse and dependence who seek treatment or help for these disorders.

The 2001-2002 NESARC is a representative sample of the adult population of the United States, including Alaska and Hawaii. As described in detail elsewhere,³² the target population was the civilian population, 18 years or older, residing in households and group quarters, including military off-base housing, boarding houses, rooming houses, nontransient hotels and motels, shelters, facilities for housing workers, college quarters, and group homes. Inclusion of the group quarters sampling frame was a strategy designed to increase representation of individuals with drug use disorders in the NESARC sample. Face-to-face interviews were conducted with 43 093 respondents. The NESARC oversampled black and Hispanic subjects and young adults (aged 18-24 years). The overall response rate was 81.0%.

The complex sampling design necessitated adjusting the data to reflect the probability of selection of primary sampling units within strata, selection of housing units within the primary sampling units, and oversampling. Adjustments for nonresponse at the household and person levels were accomplished by equating weights for the responders across predictor variables (ie, age, race or ethnicity, sex, region, poverty level, marital status, and income) with the corresponding weights of both responders and nonresponders, a standard iterative procedure used to minimize nonresponse bias in complex sample surveys. The weighted data were then adjusted to be representative of the US population for a variety of socioeconomic variables.

Diagnoses were made according to the criteria of the DSM-IV using the Alcohol Use Disorder and Associated Disabilities Interview Schedule–DSM-IV Version (AUDADIS-IV), a fully structured diagnostic interview designed for use by experienced lay interviewers.³⁴ The AUDADIS-IV separately assessed DSM-IV criteria for nicotine dependence and alcohol- and drug-specific abuse and dependence for the following 10 classes of drugs: sedatives, tranquilizers, opiates (other than heroin), stimulants, hallucinogens, cannabis, cocaine (including crack cocaine), inhalants/solvents, heroin, and other drugs. The withdrawal criterion of the drug dependence diagnoses was also drug specific and measured as a syndrome, requiring the requisite number of positive symptoms as defined in each respective DSM-IV withdrawal category. Drug-specific abuse and dependence diagnoses were aggregated to yield any drug abuse and any drug dependence diagnoses. Presentation of aggregate measures of drug abuse and drug dependence are standard in the field owing to sample size constraints. (See eTable 1 for prevalences of drug-specific disorders.)

Consistent with the DSM-IV,²⁴ lifetime diagnoses of abuse required a respondent to meet at least 1 of the 4 criteria defined for abuse in the 12-month period preceding the interview or before. The AUDADIS-IV dependence diagnoses required the respondent to satisfy at least 3 of the 7 DSM-IV criteria for dependence during the past year or before. Diagnoses of dependence before the past year were required to satisfy the time-clustering criteria defined in the DSM-IV, ie, at least 3 dependence symptoms must have occurred within the same 1-year period. Drug abuse and dependence are independent diagnoses in the DSM-IV, and abuse is not a prerequisite for dependence. Dependence diagnoses preempt diagnoses of abuse for individuals classified with both of these disorders. Thus, following DSM-IV, respondents classified with dependence included those with and without abuse, whereas abuse was reserved for those without dependence diagnoses. Alcohol abuse and dependence and nicotine dependence diagnoses in this report followed the same algorithms.

The good to excellent reliability and validity of the AUDADIS-IV drug use disorder criteria and diagnoses (κ = 0.53-0.79) are well documented in numerous psychometric studies,^{33 ,35 - 38} including clinical reappraisals conducted by psychiatrists, in clinical and general population samples,^{35 ,39} and in several countries as part of the World Health Organization/National Institutes of Health International Study on Reliability and Validity.^{40 - 45} Reliability and validity^{33 ,36 - 50} of alcohol use disorder and nicotine dependence diagnoses were also good to excellent.

It is important to note that the data presented in this study on DSM-IV drug use disorders will differ from the corresponding data derived from the 2001-2002 US National Comorbidity Survey Replication^{51 - 52} and the related World Mental Health surveys⁵³ conducted in other parts of the world. The survey instrument used in the US National Comorbidity Survey Replication, the World Mental Health–Composite International Diagnostic Interview, used drug abuse questions as screens for drug dependence, ie, respondents with no positive abuse symptoms were not asked symptom questions about drug dependence. Because a large proportion of individuals with drug dependence do not have drug abuse, cases of dependence without abuse were missed in the National Comorbidity Survey Replication. Empirical evidence^{54 - 56} has shown that the use of this screening method misses 22.4% and 8.8% of current and lifetime DSM-IV dependence cases, respectively (especially among women). Furthermore, the World Mental Health–Composite International Diagnostic Interview does not yield drug-specific abuse or dependence diagnoses unless associated problems were reported for only 1 substance. Because these limitations will alter the relationships between drug dependence and sociodemographic characteristics, disability status, treatment- or help-seeking behavior, and comorbidity, comparisons were not made between the National Comorbidity Survey Replication and related studies and the present study. In contrast, the NESARC AUDADIS-IV provides complete coverage of DSM-IV drug dependence among all individuals who ever used any drugs and also assesses drug-specific abuse and dependence.

The AUDADIS-IV assessed 5 DSM-IV anxiety disorders (panic with and without agoraphobia, social phobia, specific phobia, and generalized anxiety) and 4 major mood disorders (dysthymia, major depressive disorder, and bipolar I and II disorders). These disorders also followed DSM-IV criteria, required the clinical significance criterion to be met, and ruled out substance-induced episodes.²⁴ The AUDADIS-IV assessed the following 7 PDs on a lifetime basis: avoidant, dependent, obsessive-compulsive, paranoid, schizoid, histrionic, and antisocial. The DSM-IV PD diagnoses required evaluating long-term patterns of functioning, social/occupational impairment, and exclusion of substance-induced cases, as well as those occurring during the course of related Axis I disorders.^{55 - 58} As described in detail in ARCHIVES and elsewhere, the reliability and validity of mood, anxiety, and PDs were fair to good as assessed in clinical and general population samples.^{57 - 64} Psychotic disorders were not assessed in the NESARC and rarely are assessed in population surveys owing to their low prevalence and poor reliability and validity of the diagnoses.

Disability among respondents with drug use disorders was determined with the 12-item Short-Form Health Survey, version 2 (SF-12v2),⁶⁵ a reliable and valid impairment measure widely used in population surveys. The SF-12v2 mental impairment scales included the mental component summary, mental health, social functioning (limitations due to emotional problems), and role emotional functioning. Each SF-12v2 norm-based disability score is a continuous measure with a mean of 50 points (meaning an expected value of 50 in the general population) and a standardized range of 0 to 100 points. Lower scores indicate more disability.

Respondents were asked about drug treatment- or help-seeking behavior in the following settings: self-help groups; family/social services; drug detoxification; inpatient ward of a hospital; outpatient clinic; rehabilitation unit; methadone program; emergency department; halfway house; crisis center; employee assistance program; private physician, psychiatrist, psychologist, or social worker; counseling with a member of the clergy; and any other treatment- or help-seeking behavior.

Weighted frequencies, cross-tabulations, and means were used to derive 12-month and lifetime estimates of the prevalences of drug abuse and dependence and treatment- or help-seeking behavior among the total sample and sociodemographic subgroups. Odds ratios (ORs), derived from logistic regression analyses, indicated the associations of 12-month and lifetime DSM-IV drug abuse and dependence with sociodemographic factors, treatment- or help-seeking behavior, and comorbid disorders. Associations between drug abuse and dependence and each specific other psychiatric disorder were examined, adjusting first for only sociodemographic factors. Second, we additionally controlled for other psychiatric disorders to assess the unique relationship between drug abuse and dependence and other psychiatric disorders, which, importantly, adjusts for comorbidity of these disorders with each other. The extent of comorbidity (number of comorbid disorders) was also compared between individuals with drug use disorders in the general population and among those seeking treatment or help. Hazard rates, reflecting the risk of onset of drug abuse and dependence at specific ages among the population at risk at those ages, were calculated using standard life-table methods.⁶⁶ Disability or impairment among respondents with drug use disorders was determined using multiple regression analyses to assess the relationship between 12-month drug abuse and dependence and the 4 SF-12v2 mental disability scores, controlling for sociodemographic characteristics and all other substance use, mood, anxiety, and PDs. All standard errors and 99% confidence intervals were estimated using SUDAAN, version 9.0,⁶⁷ which adjusts for the NESARC design characteristics.

The 12-month and lifetime prevalences of drug use disorder were 2.0% and 10.3%, respectively (Table 1). Twelve-month and lifetime prevalences of drug abuse (1.4% and 7.7%, respectively) exceeded the corresponding rates for drug dependence (0.6% and 2.6%, respectively) for the total sample and virtually every sociodemographic subgroup of the population. Table 2 shows the risks of 12-month and lifetime abuse and dependence in population subgroups via adjusted ORs and 99% confidence intervals.

For 12-month disorders, the odds of drug abuse were greater among men, white compared with Hispanic respondents, those in the lowest income category, those residing in the West relative to the Midwest and South, and respondents who were never married or were widowed, separated, or divorced. The odds of dependence were greater among men, Native American compared with white respondents, those in the lowest income and education groups, and respondents who were never married or were widowed, separated, or divorced.

For lifetime abuse and dependence, the odds were higher among men, those residing in the West compared with the Northeast and Midwest, and respondents who were widowed, separated, or divorced, but lower among Asian and Hispanic compared with white respondents. The odds of lifetime abuse were also lower among black relative to white respondents and among those residing in the South. The odds of lifetime dependence were additionally greater among Native Americans, respondents in the 2 lowest income brackets, and respondents who were widowed, separated, or divorced.

The odds of 12-month and lifetime drug abuse and dependence were greater in the 3 youngest age groups relative to the oldest age group. As can be seen in Figure 1, both drug abuse and drug dependence have a similar adolescent/early adult age at onset, peaking at about 19 years, with onsets after 25 years of age quite rare. The mean age at onset of drug abuse (19.9 years) was significantly (P = .04) younger than that of dependence (21.7 years).

Mean (SE) mental component summary, mental health, social functioning, and role emotional functioning scores on the SF12-v2 for those with 12-month drug abuse were 48.8 (0.54), 48.9 (0.53), 49.3 (0.50), and 48.5 (0.56), respectively, whereas corresponding scores for those with 12-month drug dependence were 41.9 (1.15), 43.7 (0.95), 42.3 (1.10), and 43.7 (1.22), respectively. After adjusting for sociodemographic characteristics and other Axis I and II disorders, drug abuse was associated with lower mental component summary (β = −2.3; P<.001), mental health (β = −1.9; P<.001), social functioning (β = −2.1; P<.001), and role emotional functioning (β = −2.3; P<.001) scores. Drug dependence was highly associated with lower mental component summary (β = −4.3; P<.001), mental health (β = −4.2; P<.001), social functioning (β = −5.1; P<.001), and role emotional functioning (β = −3.6; P<.001) scores. Thus, although respondents with drug abuse and dependence were significantly more disabled than those respondents who did not have these disorders, dependence was clearly more disabling than abuse.

Comorbidity between DSM-IV abuse and dependence and other psychiatric disorders adjusted for sociodemographic factors is shown in Table 3 for 12-month and in eTable 2) for lifetime disorders. With few exceptions, 12-month and lifetime drug abuse and dependence were positively and significantly related to alcohol use disorders, nicotine dependence, and mood, anxiety, and PDs.

Table 3 also shows the same associations, controlling for sociodemographic factors and all other psychiatric disorders. These ORs are lower than those appearing in other NESARC publications^{59 - 64} and other previous epidemiologic surveys^{5 - 8 ,26 - 31} because they additionally control for the comorbidity of other psychiatric disorders with one another. The reductions in the magnitude and significance of the associations when the confounding effects of other psychiatric disorders were controlled for were striking. Although drug abuse remained highly and significantly associated with other substance use disorders and antisocial PD, there were no other significant associations observed between drug abuse and mood, anxiety, or PDs. The only exception was the significant but weak association between drug abuse and major depressive disorder and negative associations with paranoid and histrionic PDs on a lifetime basis. Twelve-month drug dependence remained positively and significantly related to substance use disorders and each specific mood disorder (except bipolar II disorder), generalized anxiety, and antisocial PD. Lifetime drug dependence was also associated with panic with and without agoraphobia.

Twelve-month treatment rates of drug abuse and dependence were 6.1% and 30.7%, respectively, and corresponding lifetime treatment rates were 8.1% for abuse and 37.9% for dependence. Mean ages at onset of first treatment for drug abuse and dependence were 26.7 and 27.2 years, respectively (P = .24). However, treatment rates increased significantly (P<.01) compared with treatment rates 10 years earlier, when 4.1% of respondents with 12-month abuse, 19.5% with 12-month dependence, 9.2% with lifetime abuse, and 30.1% with lifetime dependence reported having sought treatment.²³

Among those with 12-month drug abuse, 2.3% received treatment from physicians or other health care professionals; 2.0%, from self-help groups; 1.3% to 1.6%, from detoxification units, outpatient clinics, rehabilitation programs, and inpatient facilities; and 0.1% to 0.4%, from other treatment sources (eTable 3). Of those with 12-month dependence, 19.5% and 18.8% received treatment from physicians or other health care professionals and 12-step programs, respectively, with lower treatment rates for detoxification units, outpatient and inpatient facilities, and rehabilitation programs (10.0%-14.7%). Respondents with lifetime drug use disorder showed similar patterns of treatment- or help-seeking behavior by setting.

None of the sociodemographic characteristics predicted treatment for 12-month DSM-IV drug abuse and dependence. The odds of lifetime treatment for drug abuse and dependence were significantly (P<.01) greater, however, among respondents who were widowed, separated, or divorced (ORs, 1.9 and 1.8, respectively) and in the lowest income bracket (ORs, 1.9 and 2.1, respectively). Although few sociodemographic characteristics were associated with treatment- or help-seeking behavior, comorbid psychiatric disorders were strongly associated. Comorbidity was greater among those with a drug use disorder who had sought treatment or help compared with respondents with drug use disorders in the general population (Figure 2).

Our results indicate that in 2001-2002, 2.0% of adult Americans experienced a drug use disorder in the preceding 12 months (1.4%, abuse; 0.6%, dependence), whereas 10.3% developed a drug use disorder at some time during their lives (7.7%, abuse; 2.6%, dependence). Drug abuse and dependence were associated with significant disability and early onset. Thus, drug use disorders continue to be a widespread and substantial public health problem in the United States.

Rates of drug abuse and dependence were significantly greater among men than among women, a finding consistent with previous epidemiologic surveys.^{5 ,7 - 8 ,21 ,23 ,26 - 31} Age was significantly and inversely related to 12-month drug abuse and dependence, a finding also observed in earlier epidemiologic studies.^{21 - 22 ,68 - 69} However, the lifetime rates and odds among individuals in the 2 youngest age groups were nearly identical. These results indicate a potential for increases in rates for older cohorts as the generation X cohort (those aged 30 to 44 years) ages.^{70 - 71} The equally high rates among the youngest NESARC cohort who went through adolescence between 1985 and 2001 may in part be due to the rising rates of marijuana⁷² and methamphetamine^{73 - 74} use disorders observed between the 1991-1992 and 2001-2002 study periods, possibly reflecting the increased potency of each of these drugs during that decade. The near equivalence of lifetime rates between these 2 cohorts, despite the fact that individuals in the youngest age group have had shorter durations of these conditions, portends a potential epidemic in the youngest NESARC cohort. This study also found that onsets of drug abuse and dependence were typically during late adolescence or early adulthood. The implications are that adolescence is a particularly vulnerable period for the onset of drug use disorders and should be a target for etiologic and prevention research.

Rates of drug use disorders among Native Americans were not reported in national surveys before 2001,²¹ obscuring the higher rates of drug abuse and dependence in this group. This finding is consistent with those of regional studies of Native Americans.^{75 - 76} The 12-month rates of drug abuse and dependence among white respondents did not significantly differ from those of Asian or black respondents. However, that drug abuse and dependence among these minority subgroups may be increasing is reflected in their significantly lower rates compared with white respondents on a lifetime basis, relative to the lack of such differences in the 12-month rates. Although these results are suggestive of changes over time among these minority groups, this issue would best be addressed within a longitudinal framework.

Consistent with the findings of previous surveys,^{21 ,23 ,68 - 69} prevalences of DSM-IV 12-month drug abuse and dependence were generally greater among individuals with lower socioeconomic status, in terms of lower education or income levels, and among those residing in the West. The long-established Mexico-based polydrug trafficking organizations could in large part be responsible for the increased rates in the West relative to other regions of the country. Further detailed analyses of the NESARC and other similar data are needed to examine the reasons for these sociodemographic differentials within the context of drug availability, neighborhood environments, psychiatric comorbidity, and genetic predisposition toward both drug taking and drug abuse and dependence.^{77 - 85}

A new finding in this study is the importance of controlling for other psychiatric disorders (which are highly comorbid with each other) when examining associations between drug use disorders and specific psychiatric disorders. Consistent with the results of previous epidemiologic surveys,^{6 - 8 ,58 - 62 ,68 - 69} strong and significant associations were found between drug abuse and dependence and other Axis I and II disorders when we controlled for sociodemographic characteristics. To understand this comorbidity, however, we examined the unique relationships of other specific psychiatric disorders with drug use disorders, additionally controlling for the confounding effects of the comorbidity of other psychiatric disorders. Associations between drug abuse and dependence and other substance use disorders were reduced but remained strong, as were associations between drug dependence and mood and generalized anxiety disorders. Consistent with results from twin and genetic studies,^{86 - 88} the decreased magnitudes of these associations suggest that common factors may underlie these associations. Consistent with this genetically informed research, associations also remained strong, suggesting that unique factors underlie these specific drug use disorders. Taken together, these findings highlight the importance of continued research on common and specific factors underlying the comorbidity of drug abuse and dependence and these disorders.

Treatment rates for drug use disorders in the 2001-2002 compared with the 1991-1992 period⁸⁹ showed modest increases for drug abuse and dependence. Nevertheless, lifetime treatment rates for drug use disorders (abuse, 8.1%; dependence, 37.9%) are substantially lower than corresponding treatment rates of other major psychiatric disorders, eg, major depressive disorder (60%),⁶⁴ bipolar I disorder (60%),⁶³ panic disorder with agoraphobia (65%),⁶¹ and generalized anxiety disorder (50%).⁶² The lack of more significant progress in treatment for drug use disorders has been attributed to stigma,^{90 - 91} clinical lack of knowledge and uncertainty regarding screening,^{92 - 93} and insufficient organizational support.⁹⁴ There is also concern by the public, including those with drug use disorders, regarding the effectiveness and worth of available drug treatment.⁹⁴

Clearly, there is a need for a national public educational campaign to destigmatize drug use disorders and approaches to educate physicians and the public about treatment for drug use disorders. Dissemination of information on recent advancements in drug abuse treatment,^{95 - 98} including manual-driven, empirically validated treatment approaches to reduce the use of a variety of drugs (eg, motivational enhancement, 12-step, and/or cognitive-behavioral therapies) and new medications to combat drug craving and withdrawal symptoms of some substances (eg, heroin and cocaine), is required.^{92 ,99}

The most common treatment settings for individuals with drug use disorders included private physicians and other health care professionals, a finding that underscores the continued importance of the critical detection and referral roles of primary care physicians in the treatment of these disorders. Future research efforts focused in primary care settings on the development of instruments to screen, identify, and refer probable abuse and dependence in primary care settings, similar to the National Institute on Alcohol Abuse and Alcoholism guide to the identification of patients with alcohol use disorders,¹⁰⁰ appear warranted, as do efforts to computerize assessment and referral processes.¹⁰¹ Because not all individuals with drug problems (eg, some individuals with drug abuse) require formal treatment, skills training in motivational interviewing and brief interventions should also become a standard in clinical training curricula. The core features of these psychosocial interventions apply across the full range of drug use disorders, with modifications unique to particular categories of drugs.^{102 - 104} Furthermore, validated screening instruments and training are also needed for agencies in key positions to screen, assess, and refer those with drug use disorders (eg, criminal justice and welfare) by virtue of the high frequency of these disorders in their populations.^{94 ,105}

Limitations of this study include its cross-sectional nature, and several issues addressed herein would best be examined in a longitudinal context. The wave 2 NESARC, a 3-year follow-up of participants in this wave 1 survey, was designed to address this limitation and to provide a strong platform to further investigate the stability of the observed relationships in the general population. Although test-retest reliability and validation through clinical reappraisal studies conducted by physicians speak to the reliability and validity of the drug use disorder diagnoses presented herein, some degree of underreporting of illicit drug use and symptoms is likely in all surveys of the general population when self-report assessment instruments are used. Furthermore, general population surveys may fail to capture all individuals with drug use disorders because these individuals are less likely to live in households, the exclusive sample frame of most general population surveys. However, as previously discussed, the NESARC sampled from households and group quarters (eg, shelters and group homes), a strategy designed to increase the representation of individuals with drug use disorders in the sample. Based on these considerations of potential underreporting and underrepresentation of individuals with drug use disorders, NESARC estimates of prevalence, risk, and comorbidity are likely to be conservative.

In summary, the NESARC has shown that DSM-IV drug abuse and dependence are prevalent, highly disabling disorders that often go untreated. Drug use disorders, especially drug dependence, are highly comorbid, highlighting the need for comprehensive assessment and treatment of comorbid disorders. The study identified population subgroups at particular risk and generated many findings that can lead to further hypothesis-driven investigations. The adolescent onset of drug abuse and dependence revealed critical windows of opportunity for prevention efforts. The results of this study indicate that immediate action must be taken to educate physicians, the public, and policy makers about drug use disorders and their treatment and to develop programs to destigmatize the disorders, thereby reducing the personal suffering and adverse societal impact of drug use disorders in the United States.

Correspondence: Bridget F. Grant, PhD, PhD, Laboratory of Epidemiology and Biometry, Room 3077, Division of Intramural Clinical and Biological Research, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Mailstop 9304, 5635 Fishers Ln, Bethesda, MD 20892-9304 (bgrant@willco.niaaa.nih.gov).

Submitted for Publication: May 10, 2006; final revision received August 25, 2006; accepted October 6, 2006.

Financial Disclosure: None reported.

Funding/Support: This study was supported in part by the Intramural Program of the National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism. The NESARC is funded by the National Institute on Alcohol Abuse and Alcoholism, with supplemental support from the National Institute on Drug Abuse.

Disclaimer: The views and opinions expressed in this report are those of the authors and should not be construed to represent the views of any of the sponsoring organizations, agencies, or the US government.

Additional Information:eTable 1, eTable 2, and eTable 3 are available.