Understanding Transitions in Illicit Drug Use and Drug Use Disorders

One reason for our modest success in the prevention and treatment of drug use and drug use disorders is our limited knowledge of their natural history. A better understanding of their complex natural course could advance our apprehension of their etiology and inform efforts to develop more effective interventions. The work of Compton et al. (1) presented in this issue of the Journal represents an important step in this direction.

Compton et al. examine transitions in drug use status among individuals who participated in the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). The NESARC, which was sponsored by the National Institute on Alcohol Abuse and Alcoholism, studied a large, nationally representative sample of U.S. adults. It collected extensive information on drug use, drug use disorders and other psychiatric disorders (including all DSM-IV personality disorders), and a broad range of potential predictors of the course of psychiatric disorders. Notably, NESARC participants underwent two assessments, 3 years apart, allowing for the prospective study of the course of drug use transitions. Compton et al. examine all potential transitions among three situations: nonuse, asymptomatic use, and problem use. The authors adopt a dimensional view of drug problems by focusing on problem use rather than on the more common categories of drug abuse and dependence, consistent with more recent conceptualizations of psychopathology (2, 3).

The study had two major findings. The first was the documentation of the different courses that drug use can follow over time in the general population, an important piece of information that, surprisingly, was unknown until now. Among nonusers at baseline, less than 5% started using drugs in the 3 years between the two assessments, but of those who started using, more than half were experiencing problems at the follow-up assessment. Two-thirds of users who were asymptomatic at baseline had stopped using at follow-up, whereas almost another 20% had progressed to problem use. Of those with problem use at baseline, almost half (49%) were not using 3 years later, and another 40% continued having problems (the other 11% had transitioned to asymptomatic use).

The second major finding was that these transitions in drug status are influenced by a variety of factors, which differ depending on the type of transition. The investigators were able to identify more predictors of drug use initiation than of drug use cessation or progression to asymptomatic use. For example, among participants who were nonusers at baseline, younger age, alcohol and nicotine use, and major depressive disorder were associated with drug use initiation by the time of the follow-up assessment. In this group, lack of health insurance, poor physical and psychological health, and a family history of substance use disorders were associated with a higher risk of transitioning from drug initiation to problem use rather than to asymptomatic use. The risk of transition from asymptomatic to problem use was higher among those with poverty or borderline personality disorder and lower among those with a cluster C personality disorder. A history of child abuse, early drug use, and having schizotypal or narcissistic personality disorder was associated with a higher risk of persistent problem drug use.

These findings, which have high generalizability because of the national representativeness of the sample, have important implications in several domains. For clinicians, these results converge with previous analyses in indicating that most patients with drug use disorders are likely to improve and to stop substance use (4). They provide some good reasons to be optimistic regarding the course of individuals with drug use disorders. At the same time, a substantial proportion of study subjects with problem use at baseline continued to have problem use at the 3-year follow-up, many asymptomatic users progressed to problem use, and the majority of nonusers who initiated drug use after the baseline assessment had developed problem use by the time of the follow-up assessment. These results provide support for the view of drug problems as potentially chronic conditions that require ongoing monitoring and management (5). Taken together, the findings offer a mixture of optimism and caution concerning treatment of individuals with drug problems.

From the epidemiological perspective, a major contribution of this study is the demonstration that the course of drug use and problem use is influenced by risk factors at multiple levels, including sociodemographic characteristics, history of childhood abuse, family history of substance use disorders, current health status, use of other substances, and axis I and II psychopathology. Given this complexity, treatment and preventive interventions are likely to be most successful if they intervene on more than one risk factor. An important direction for future research will be the development of comprehensive development models of the etiology and course of drug use disorders, similar to those that already exist for depression (6). These models, by allowing us to understand how risk factors interact with each other over time, may help guide the development of age- or development-specific interventions. For example, prevention and treatment of childhood abuse or close monitoring of individuals with a family history of drug use disorders may be particularly important intervention targets for children and adolescents, whereas treating comorbid psychiatric disorders may be more important in adults. The interventions may also have to be different depending on whether they seek to address drug initiation, transition to problem use, or treatment of drug problems.

Three study findings had particular implications for prevention and policy. Younger age at the time of interview and early onset of drug use were associated with a higher risk of transitioning from nonuse to problem use, emphasizing the need for prevention and early intervention among youths. The finding that more than half of individuals who initiated drug use after the baseline evaluation developed problem use within 3 years of their onset of use is a powerful reminder of the addictive power of drugs. It suggests caution regarding the potential unintended consequences of policies that may increase the availability of drugs based on limited scientific evidence (7). The third finding, that poverty and lack of health insurance predict transition from nonuse to asymptomatic use to problem use or decrease the probability of transition to nonuse is also an important reminder of the disproportionate burden of drug problems experienced by the socioeconomically disadvantaged, and of the hope that greater access to treatment will contribute to decreasing this burden (8, 9).

One lesson about of the approach of the study also deserves mention. The investigators were able to address these vitally important questions with existing data, making efficient use of readily accessible resources and dramatically decreasing the cost and time needed to conduct their study. There are currently many publicly available data sets that can be used to investigate important epidemiological and clinical questions, and in the coming years additional data sets will become available to address many other aspects of psychiatric disorders. Developing the necessary funding streams and training resources to fully benefit from these data sets may offer unprecedented opportunities to scientists at all levels to dramatically accelerate our knowledge of the etiology and treatment of drug use and other psychiatric disorders (10).