Findings and Research Recommendations
Over the past 25 years, governments at all levels have made great efforts to reduce drug use in the United States. Government expenditures on drug control total about $40 billion annually. The number of people incarcerated for drug offenses on any given day has risen from less than 50,000 in the early 1980s to about 500,000 in 2005, including many held in local jails (Caulkins and Chandler, 2006). Treatment has expanded substantially, and much has been learned about what constitutes effective treatment. And although the prevalence of illegal use is below the highest levels achieved in the late 1970s (for marijuana) and in the mid-1980s (for cocaine), it has remained stubbornly high. For many young people today, occasional marijuana use is a part of adolescent development, as it has been since the birth cohort of about 1960. At the same time use of diverted prescription drugs has been rising.
The data available for understanding why the nation still has a large demand for illegal drugs is woefully inadequate. For example, even the most basic numbers for policy purposes—estimates of the number of chronic users of cocaine, heroin, and methamphetamine; the quantity used; and the amount spent purchasing them—have not been published for almost a decade (Office of National Drug Control Policy, 2001). In this chapter we offer recommendations for data collection and research activities, both epidemiology and treatment research, that might enable the government to respond effectively to the continued demand for illegal drugs.
PREVIOUS RECOMMENDATIONS AND CURRENT DATA USE
We start by noting that the report of a prior committee of the National Research Council (2001) made numerous recommendations for strengthening the federal data collection effort. That committee concluded, for example, that there was a dearth of data on consumption (as opposed to prevalence) and that this topic deserved high priority. The committee expressed concern that access to critical prevalence datasets was unreasonably restricted. It recommended, for example, that the Substance Abuse and Mental Health Services Administration (SAMHSA) work out arrangements along the lines developed by the National Center for Educational Statistics to allow researchers to work with restricted datasets in a way that preserves confidentiality but still allows full utilization of the data. It also recommended that the National Institute on Drug Abuse (NIDA) ensure that the longitudinal panels of Monitoring the Future (MTF) Survey become available for outside researchers.
Although there have been some improvements in data systems in recent years, such as the National Survey on Drug Use and Health (NSDUH) and the Treatment Episode Data Set (TEDS), few of the recommendations of the 2001 committee report have been implemented. In this report we repeat and emphasize some of those previous recommendations, and we supplement them by providing specific comments on individual datasets that are critical to assessing the determinants of the demand for illegal drugs. We note that we did not have the resources to develop a full research agenda for assessing the effectiveness of existing programs and policies aimed at reducing demand. Before providing recommendations on data and research, we note that existing data are often misused. Advocates, legislators, and policy makers frequently cite level and trend data from the NSDUH or other survey data sources that are misleading. In some cases, such data are used to produce point estimates of some quantities, such as annual U.S. cocaine consumption, that are known to severely understate the true figures. In other cases, comparisons are made about drug use behaviors among sociodemographic groups with very different response rates and that do not have internal validity.
It is important that policy makers and the public be aware of systematic limitations and potential biases when these data are used. In addition, it would be helpful to policy makers and analysts if the Office of National Drug Control Policy (ONDCP), the Centers for Disease Control and Prevention, and other agencies would provide more extensive, explicit, and accessible guidance regarding the known limitations of existing data that might otherwise be miscommunicated or misused.
RECOMMENDATIONS FOR DATA SYSTEMS
National Survey on Drug Use and Health
NSDUH is now one of the largest annual surveys of the household population conducted by the federal government, with approximately 67,000 respondents in 2007, and it represents several methodological improvements over its predecessor, the National Household Survey of Drug Abuse (NHSDA). These improvements have increased response rates and likely also improved data quality. NSDUH remains a critical data source in understanding trends and correlates associated with illegal drug use.
Given the importance of these data, it is a high priority to improve NSDUH’s utility for both policy making and research. Some of these improvements reflect the need to implement recommendations offered in the previous National Research Council (2001) report, Informing America’s Policy on Illegal Drugs: What We Don’t Know Keeps Hurting Us. We endorse those recommendations, many of which still remain pressing.
In particular, we echo the previous committee’s recommendations on data collection and systematic research:
The committee recommends that the National Survey on Drug Use and Health (NSDUH) and other national surveys expand their data collection efforts to more effectively survey subpopulations with high prevalence of substance use. In the specific case of NSDUH, we recommend that methods be developed to survey the institutionalized populations that are currently excluded from the data.
The committee recommends that the National Institute on Drug Abuse, the Office of National Drug Control Policy, and other funding organizations support a systematic research program (a) to understand the effects of nonresponse and reporting errors in surveys and (b) to design and implement surveys in ways that minimize the resulting biases.
It is especially critical that this research include systematic work on the strengths, weaknesses, and best practices of increasingly widespread methodologies used to reach and survey hidden populations. Research to improve and scrutinize such methods as respondent-driven sampling deserve priority, given that nontraditional survey methodologies are required to reach such hidden populations as street-injection drug users.
Methodological research on the impact of respondent incentive payments on research participation, for example, would be quite valuable.
Although we reaffirm the earlier recommendations regarding NSDUH, the committee finds that this important survey and others used to monitor substance abuse have some important weaknesses from a policy-making perspective.
NSDUH was designed and is mainly used to provide descriptive information on basic trends and correlates of illegal drug use. Both policy analysis and social science research require the ability to systematically link such data with a rich set of variables regarding personal circumstances, public policies, and individual encounters with social service systems. NSDUH (and its predecessor, NHSDA) have operationalized key variables in ways that undermine comparability with other available data and that are not always consistent over time.
The committee recommends that the National Survey of Drug Use and Health follow current best practices methodologies and be more systematically and explicitly coordinated with other high-quality datasets in areas important to substance abuse.
One example of such a dataset is the National Comorbidity Study and its successors, which provide valuable instruments in screening for psychiatric disorders and in understanding barriers to the treatment of such conditions. Two other examples are the Panel Study of Income Dynamics and the Women’s Employment Study, which provide valuable information on the receipt of food stamps, support from Temporary Assistance to Needy Families, and other forms of public aid. Yet another valuable dataset is the National Health Interview Survey, which is widely used in substance abuse policy research.
Many of the most important policy research questions concern public policies and other factors that operate at the state level. Analyses at that level are typically hindered by the common practice of masking geographic identifiers, even in datasets for which such geographic identifiers pose little risk of respondent disclosure. Procedures to secure access to these identifiers are often arbitrary or clumsy, parallel to those required to gain access to much more confidential data. The policies for these datasets are often more restrictive than those that are generally applied to census and birth record data, which include confidential data about a much larger fraction of the American public.
The committee recommends that individual-level survey data released for public or research use should routinely include state
identifiers barring specific justification that such identifiers pose a significant risk of respondent disclosure.
We are encouraged that SAMHSA is developing procedures to facilitate researcher access to data with substate identifiers, and we hope that SAMHSA’s Office of Applied Studies will move this effort forward.
Large, nationally representative datasets such as NSDUH are designed to provide reasonably precise estimates of key parameters, such as the prevalence of recent illicit drug use. Although these data are valuable to policy makers, the large sample size and national coverage comes at some price. These datasets do not provide extensive coverage of key populations, such as people in rural areas and homeless people. Moreover, these national datasets do not provide the descriptive detail required to scrutinize the interactions among diverse drug users and sellers in drug markets. For this reason and others, nationally representative datasets must be complemented with more extensive localized studies, including qualitative and ethnographic studies and studies that use administrative data, to provide a more granulated view than can be obtained from national surveys.
Other available datasets, collected for different purposes, could provide improved guidance for policy makers and researchers on the changing epidemiology and market conditions pertinent to illegal drug use. National surveys of treatment facilities, including TEDS, the National Drug Abuse Treatment System Survey (NDATSS), and the National Survey of Substance Abuse Treatment Services, were mainly developed and designed to explore service delivery issues in substance abuse treatment. These data can be augmented to include richer data on clients’ drug use behaviors and treatment outcomes. For example, NDATSS and TEDS include rich data regarding the primary drug of abuse and sociodemographic characteristics of clients entering treatment facilities. We caution, however, that these data have to be interpreted with care because they are not representative of the full population of drug users. By definition, these data exclude people who do not enter treatment, but they remain useful for many purposes.
These datasets typically also include a larger number of drug-dependent people with particular circumstances and disorders than is available through NSDUH. For example, the 2006 TEDS included admissions data for more than 22,000 pregnant women entering substance abuse treatment. In contrast, the 2006 NSDUH included less than 200 pregnant women who reported using any illegal substance in the previous year.
Unit-level datasets might also provide sample frames for individual-level surveys. National cohort studies, such as the National Treatment Improvement Study, have provided valuable information for policy makers on drug use careers and other matters. Although individual-level data are available from many sources, there has been no comparable national cohort study in more than a decade.
Monitoring the Future Survey
The MTF survey of approximately 50,000 respondents has been collecting data for more than 35 years with great consistency of questions and with results reported in a timely fashion. It is conducted by the Survey Research Center at the University of Michigan, with funding from NIDA at the National Institutes of Health. For the past 8 years, it has been the principal indicator for the federal government in assessing the success of the federal drug strategy in reducing illegal drug use, particularly for adolescent substance use. However, its value has been limited by not adding noncore items that would allow a better understanding of changes in drug use among youth. The usefulness of the MTF data would also be enhanced if outside researchers had better access, with appropriate safeguards, to information necessary to explore methodological issues about survey design. For example, each year 30-50 percent of the schools selected to participate in the MTF survey decline to participate (for details, see National Research Council, 2001). Research is needed to explore the reasons for nonparticipation and the implications of the method used to find replacement schools.
MTF and other datasets, such as the National Longitudinal Survey of Youth and the National Longitudinal Study of Adolescent Health, provide especially strong survey data on adolescents and young adults. Continuing these cohorts over time offers a valuable opportunity to explore drug use careers in the age ranges that are less often studied as respondents reach adulthood.
However, the longitudinal panels of MTF still are not used extensively by researchers. There are now 35 distinct panels, some with observations extending over decades, and they could provide immensely valuable information for understanding the dynamics of drug use careers in the general population. We are concerned that the research community may not be fully aware of the possibility of gaining access.
The National Research Council (2001, pp. 82-84), almost 10 years ago, critiqued a lack of transparency in the MTF data-sharing plan. At present, selected portions of the MTF cross-sectional data are being made available through the Substance Abuse and Mental Health Data Archive (SAMHDA), which has been operated since 1995 by the Inter-university
Consortium for Political and Social Research (ICPSR).1 There also is an ICPSR/SAMHDA web page that introduces the idea of a process through which sharing of longitudinal data might be achieved by outside (non-MTF) principal investigators.2 These positive developments are consistent with a recently stated MTF research objective: “Objective 11: To continue to facilitate the use of the MTF databases by others—including investigators in a variety of substantive and disciplinary fields—while adequately protecting the confidentiality of the study’s many respondents” (quoted in Johnson et al., 2006).
However, the possibility of obtaining access to the MTF panel data is only mentioned on the ICPSR/SAMHDA web page cited above in an answer to a Frequently Asked Question (Is the longitudinal data available for Monitoring the Future?). This answer, reached by clicking through from the ICPSR home page to the SAMHDA home page to “Tutorials and FAQs” to “Series-Specific Questions,” is the following:
All data for a particular individual are linked (or, in the case of form-specific items, capable of being linked) in the panel dataset. The sheer amount of information greatly increases the risk of breaching confidentiality. Thus, based on policies approved by our funding source and IRB, the panel data set cannot be made available to the public in totality and without modification.
Special data requests can be made through the Web site email address. Once we get a request, information about policies and procedures is sent out. Requests are considered on a case-by-case basis, and may be fulfilled—at requestor’s cost—typically by providing data analytic access. … To make a request for this data and for further information, please contact MTF staff at: MTFinfo@isr.umich.edu.
No other website provides information about the possibility of accessing the MTF panel data.
The committee contacted the staff of the Survey Research Center (SRC) and requested information on the number of researchers who have been provided access to the MTF panel data. They reported that, between August 1997 and July 2010, they received 35 requests for access to panel data that would be used for in-depth analysis. All of these requests were granted. The SRC staff acknowledged that it is difficult to find information about their data sharing policies and procedures on the MTF website, but they explained their concern that the success of MTF’s scientific
See http://www.icpsr.umich.edu/icpsrweb/SAMHDA/studies?q=Monitoring+the+Future [accessed July 2010].
See http://www.icpsr.umich.edu/icpsrweb/SAMHDA/support/faqs/0039 [accessed July 2010].
mission of gathering data on sensitive topics depends on their promise of strict confidentiality. The staff said that they are developing a separate website regarding data sharing that will be announced soon.
Given the unique value of these MTF panel data, the committee endorses the recommendation of the previous National Research Council (2001) committee that NIDA create an advisory panel to work with the grantee on issues of data access, in particular to foster additional awareness about methods of providing access that will meet the needs of the research field.
Public Health Surveillance
The emergence of HIV/AIDS among injection drug users three decades ago, the crack epidemic of the 1980s and early 1990s, and the 2007-2008 epidemic of fentanyl-related opiate overdoses underscore the serious public health challenges associated with illegal drug use. These events also underscore the value of early warnings about emerging drug use “epidemics” and other emerging threats to the health and well-being of substance users. Our recommendation on this issue is consistent with that of the previous National Research Council (2001) committee.
The committee recommends that the Office of National Drug Control Policy, in consultation with pertinent agencies of the National Institutes of Health and the Centers for Disease Control and Prevention, develop procedures for effective epidemiological surveillance concerning emerging forms of substance use and their related harms to human health and well-being.
There have been major losses of indicator systems in recent years. In 2003, the Arrestee Drug Abuse Monitoring (ADAM) system, supported by the National Institute of Justice (NIJ) was ended, thus eliminating the most useful platform for the study of criminally active drug offenders. As noted in Chapter 2, ONDCP has restarted a much smaller version of ADAM, but little is known about it. At about the same time, the redesign of the Drug Abuse Warning Network (DAWN), which is sponsored by SAMHSA, substantially reduced its utility for research purposes.
Arrestee Drug Abuse Monitoring
Funding for the ADAM program was provided exclusively by the NIJ, the research arm of the Department of Justice, and it accounted for a large share of the NIJ budget in 2003. After lengthy consideration, the
agency decided that without support from other agencies, it was not central enough to the NIJ mission to continue. Moreover, NIJ was not responsible for funding any other statistical series, a function more regularly associated in the department’s Bureau of Justice Statistics.
ADAM is not an ideal system to study frequent drug users. Although many drug users do come in contact with the criminal justice system, treatment data show clearly that a substantial fraction of the people admitted with cocaine or heroin involvement have had no contact with that system. Nonetheless, given the centrality of ADAM for understanding the behavior of drug markets, the loss of ADAM has been a serious loss to understanding drug demand. We are not in a position to make recommendations as to the size, structure, and scope of an ADAM-like program, but we conclude that it is important to have some regular method for surveying this population.
The committee recommends that the U.S. Department of Justice reinstitute an Arrestee Drug Abuse Monitoring-like survey to collect data on the behavior of criminally involved drug users.
An ADAM-like survey is less important for the data it provides on the levels of drug use among arrestees than it is as a platform for studying the behavior of the population that accounts for most of the cocaine, heroin, and methamphetamine used in the United States. At present it appears impossible to develop estimates of the quantities used and the expenditures on illegal drugs without data from these populations.
Drug Abuse Warning Network
The redesign of DAWN in 2002, though intended to strengthen the system for collecting data on emergency department visits and on deaths related to drug use, has instead resulted in a substantial weakening of the system. Fewer hospitals are willing to participate than previously in part because of greater concerns about privacy protections and potential liability from misuse of data that have developed in recent years, particularly since the passage of the Health Insurance Portability and Accountability Act. It may be impossible to overcome this reluctance by hospitals to participate and therefore to obtain the data.
DAWN, even before 2002, played only a modest role in research on the demand for drugs (see Caulkins, 2001; Dave, 2006). It has been used occasionally, for example, as a method for calculating the relative sizes of drug-using populations in particular cities (e.g. Office of National Drug Control Policy, 2001). As noted in Chapter 3, the redesigned DAWN has yet to produce some of the basic series. Moreover, DAWN was budgeted
at approximately $17 million in 2009, which is relatively expensive for drug data indicators.
The committee recommends that the U.S. Department of Health and Human Services consider, first, whether the Drug Abuse Warning Network series should be continued given that these data have not had much value for either policy or research, and, second, if it is continued, whether it should be conducted by agencies that are already collecting information from emergency departments and medical examiners.
On the second point, for example, the Food and Drug Administration and other agencies collect data from emergency departments for other purposes, and the Centers for Disease Control and Prevention regularly receives data from medical examiners on vital statistics. It may be more useful to merge DAWN into those other systems, reducing the reporting burden and taking advantage of the expertise these other agencies have in dealing with emergency departments and medical examiners. We believe that continuing DAWN in this way would lower its cost.
Treatment Episode Data System
Although TEDS was originally developed as an administrative dataset, as noted above it has become an increasingly valuable research tool in understanding changing patterns of illegal drug use. The quality of the data appears to have improved in recent years, and it provides a large-sample dataset about the circumstances of people who enter substance abuse treatment. The publication of the TEDS-D discharge dataset, starting in 2008, is potentially a major addition to the capacity for both monitoring and studying drug abuse treatment.
TEDS also might be made more valuable through the inclusion of additional data, such as individual identifiers, including specific sociodemographic identifiers, and through additional surveys conducted on a subsample of treatment clients. Because TEDS imposes data collection burdens on providers and clients, we are not in a position to specify these changes in detail.
The Committee recommends that the Substance Abuse and Mental Health Services Administration give high priority to expanding and improving the Treatment Episode Data Set.
RECOMMENDATIONS FOR RESEARCH
There are many components to a research agenda on the demand for illegal drugs. We focus here on a particularly important, promising, and neglected area, namely, longitudinal research, which can help answer many of the most significant questions about demand.
Cohort and Longitudinal Studies
TEDS and other data sources may provide a promising sample frame for additional cohort studies. Although cohort studies conducted on the population in treatment have inherent limitations, these datasets also provide unique information on other factors, including desistence rates among chronic users and mortality and incarceration rates. For example, TEDS includes detailed information on the age of onset of drug use of admitted patients. Such data can be used to assemble panels of individuals to explore drug use careers more extensively than is currently possible.
The need for such data is increasingly pressing, given the absence of recent cohort studies comparable to prior efforts, such as the National Treatment Improvement and Evaluation Study. The earlier prominent cohort studies are more than a decade old and thus do not address some important contemporary questions, including trajectories of drug use and offending among methamphetamine users and changing patterns of marijuana and prescription drug abuse.
Other research efforts provide equally promising opportunities for informative cohort studies. For example, most academic research conducted in the Clinical Trials Network examines short-term clinical outcomes subsequent to treatment interventions. Long-term follow-up of patients in that network may provide unique opportunities to gauge long-term outcomes, as well as to collect longitudinal data on an accessible cohort of current and former drug users.
The committee recommends that the National Institute on Drug Abuse and other responsible funding agencies pursue available opportunities for cohort studies facilitated by recent research efforts.
Longitudinal epidemiological studies of drug use can be exploited more fully to inform scientific questions about the demand for illegal drugs. They can uniquely provide data for two purposes: (1) to generate estimates of the number of drug users, the quantity of a drug used, and the price paid for a drug; and (2) to track the process of initiation and desistence of drug involvement. Longitudinal studies can shed light on
the correlates, the suspected causal determinants, and the consequences of drug abuse.
There are many longitudinal studies of large representative cohorts that have been followed from childhood to adulthood in specific areas of the United States, with good retention rates: they include the Denver, Pittsburgh, and Rochester Youth Studies, the Great Smoky Mountains Study, the Iowa Youth and Families Project, the Oregon Social Learning Study, the Children-in-the-Community Study of upstate New York, and the Baltimore Prevention Research Center cohorts. There are also a few nationwide studies with repeated measures, including the National Longitudinal Study of Adolescent Health and the selected subsamples of college-aged students followed after high school in the MTF survey.
The National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) has excellent potential for contributions in this area, if its interview protocol is amended for the purposes of modeling drug demand nationwide. An additional NESARC wave of assessment is slated to occur sometime within the next 5 years, which presents an opportunity for such changes. However, that possibility needs to be considered in relation to some concerns that have been expressed about the accuracy and completeness of NESARC estimates on drug use (see, e.g., Grucza et al., 2007) and its constrained information value with respect to prescription painkillers that recently have become more prominent in drug demand research (see, e.g., Boyd and McCabe, 2008).
Estimating the Demand for Drugs
As detailed in Chapter 1, the information needed to model the demand for drugs is missing, with a heavy reliance, to date, on cross-sectional survey designs used to answer questions—for example, How many active drug users are there? What quantity of a drug are these users buying and consuming? What price are these users paying?—that could be added to ongoing longitudinal cohort studies. Virtually all existing longitudinal studies already ask participants about illegal, nonmedical, and extra-medical drug use in some depth. However, to date, the questions have been designed for purposes other than understanding the demand for drugs. These purposes include estimating age of drug use initiation for developmental research, diagnosing drug use disorders for psychiatric research, and assessing self-reports of illegal drug possession and selling for criminological research.
The committee recommends that researchers working with ongoing longitudinal surveys be encouraged to add standardized
items to collect data more useful for economic and drug demand modelling.
Such topics include prices and quantities consumed by active users. But, as discussed in Chapter 2, collecting useful measures of drug prices is difficult in markets characterized by a standardized price, such as a “dime bag” of unknown weight and purity. Additional questions, such as how far users travel to obtain drugs, could be asked to help anticipate whether displacement should be expected if price changes in one location but not all locations.
Longitudinal cohort studies offer certain advantages for collecting such data. Studies that began to interview their participants during childhoods in the 1970s, 1980s, and 1990s and have now followed them into adulthood offer enhanced validity of self-reports, because individuals interviewed repeatedly about drug use learn that they can trust the confidentiality guarantee and become unusually willing to provide frank reports. For example, the Dunedin Longitudinal Study compared experienced longitudinal cohort members with matched adult respondents: 30 percent more of the experienced subjects revealed daily cannabis use (Moffitt et al., 2010). Moreover, the recall failure that compromises validity of self-reports in retrospective surveys can be avoided if longitudinal studies reinterview participants frequently enough so that responses cover a recent and relatively brief period of time. One comparison found that the prevalence of cannabis dependence from age 18 to 32 years was doubled in prospective longitudinal study data in comparison with retrospective survey reports (Moffitt et al., 2010).
Challenges can be anticipated. New methods of questioning have to be developed to obtain reliable and valid self-reports of quantity and price. Longitudinal twin studies (such as the Virginia Twin Study of Adolescent Development or the Minnesota Twin Study) are particularly useful for evaluating the reliability of self-reports, as they provide two same-age knowledgeable informants about the illegal drug use of each participant (twin and co-twin). There is another challenge if nationwide data are needed. Existing longitudinal cohort studies whose participants have reached adulthood with good retention rates tend to represent specific cities or states. Good longitudinal sample retention is essential for modelling drug demand, because substance use is associated with nonresponse, and high rates of nonresponse characterize many contemporary surveys and make them problematic for ascertaining drug demand (see e.g., Galea and Tracy, 2007). Few longitudinal studies with good retention rates have nationwide samples.
The previous National Research Council (2001) report contained a critical discussion of the System to Retrieve Information from Drug Evi-
dence (STRIDE) data on illegal drug prices. Although we were unable to address this issue in depth, we support on-going research efforts to further assess the reliability of the STRIDE data and consider alternatives.
Tracking the Process and Correlates of Desistence
Longitudinal cohort studies have yielded many insights about the process of the initiation of drug use (and the reasons for that initiation). Over time, these studies have become well positioned to examine the process of naturalistic desistence from drug use (and reasons for desistance). Such research, if undertaken, would parallel the scientific benefits accrued from criminology research that characterized crime careers by following individual offenders over time (see, e.g., Laub and Sampson, 2003). Results from research into the predictors and correlates of naturalistic drug desistence might be harnessed to develop novel prevention and treatment approaches. Representative birth or school cohorts whose subjects have now reached their 30s are well suited to studying desistence from drug use in the general population. Longitudinal follow-up studies of adult samples can also be highly informative on other issues. For example, research should track desistence in longitudinal studies of adults sampled for their dependence on particular drugs to determine whether desistence processes vary by drug type. Long-term longitudinal follow-up of randomized clinical trials of drug treatment programs could be used to identify factors that discriminate between recovery and relapse.
Longitudinal cohort studies offer certain advantages for studying desistence. As in research on criminal careers, self-report data from longitudinal cohorts are particularly valuable for studying desistence. The alternative is official record data, which are not well suited to this purpose because drug abuse treatment records tap only a small percentage of drug users (just as official crime conviction records cover only a small percentage of offenders). Similarly, a drug-dependent person who no longer appears in treatment records after some time cannot be presumed to have ceased drug use (just as a criminal offender who no longer appears in conviction records cannot be presumed to have ceased law breaking). Another advantage is that longitudinal designs allow not only the study of drug desistence, but also the study of factors that accompany desistence and may cause it. Studies of within-individual change can take one step toward building an evidence base on causal factors by using individual drug users as their own controls, linking decreasing drug use to antecedent events while all other characteristics of the drug users remain constant. Longitudinal studies of adult twins (such as the Vietnam-Era Twin Study) can go even further toward documenting causality by exam-
ining factors that account for differences in drug use cessation between twins in monozygotic pairs.
Virtually all longitudinal cohort studies in the United States have collected data on their participants’ drug use, and many of the participants in these studies will soon reach the ages when desistence from drug use occurs. However, the capacity of these studies to follow their cohorts further for several years to cover the period of drug desistence is uncertain because they have largely been funded by federal agencies who are now under pressure to support translational research and randomized clinical trials in the context of stagnant budgets. In this funding climate, agencies may view continuing longstanding longitudinal epidemiological studies as lower priority than other research.
The committee recommends that funding agencies provide continued support for key prospective longitudinal studies that are best suited for tracking the causes and correlates of drug use desistence, including treatment and criminal justice involvement.
Finally, such panel studies as the MTF survey have repeatedly interviewed the same respondents across years, yet most of the reports from these studies concern the cross-sectional prevalence of drug use in a particular year or changes in this aggregate-level prevalence across a series of years. We have not found any MTF longitudinal analyses that inform economic or drug demand models, although the data are a national resource that could be exploited to track within-individual change in drug use over long-term study periods and to identify the correlates and suspected causal determinants of such change. Even if the MTF panels are not accessible to outside researchers, we urge grantee research groups to explore these data more deeply.
The short-term efficacy and effectiveness of various modalities of drug abuse treatment has been repeatedly and convincingly demonstrated. However, the committee notes that additional data collection and analytic approaches are needed in order to better understand how participation in treatment affects the long-term trajectory of drug use and desistence among various types of drug users. The answer to this question is needed in order to put treatment effectiveness into proper perspective in light of emerging information about the role of “spontaneous,” or untreated, recovery from drug and alcohol use. Longitudinal data would help to answer questions about the role of treatment in drug use trajectories
while taking into account other important environmental factors, such as contact with the criminal justice system.
If treatments can be effective in moving the trajectory of drug use toward earlier or more frequent desistence, then the key question for policy makers is what steps should be taken to expand and improve treatment options and programs and to increase the utilization of treatments known to be effective.
Treatment capacity is one important part of the equation, with evidence that current capacity is seriously inadequate if all drug users in need were to seek help. However, efforts to increase the acceptability and utilization of treatment may also be needed. The criminal justice system appears especially well positioned to exert a positive impact on treatment utilization and outcome through drug courts and other diversion initiatives (see Chapter 4). Again, research on the effectiveness of such innovations will be critical in order to gauge their effects on drug use trajectories and demand reduction.
There are a number of innovations and improvement that could be made to further enhance the reach and effectiveness of the treatment as currently provided. The committee is reasonably confident that treatment innovations and improvements of existing treatments will continue to emerge, in part through successful dissemination and adoption of new treatment practices. The goal of such improvements in the quality of treatment is to make drug desistence a more certain and reliable outcome of a given treatment episode or of cumulative treatment episodes which will, in turn, stimulate further demand reduction.
The committee recommends that drug abuse treatment providers adopt new practices with research conducted to examine effectiveness.
The available evidence suggests that investment in treatment capacity and other strategies of increasing participation, especially by users of opiate and stimulant drugs, would reduce the demand for drugs while generating other significant social benefits (e.g., reduced crime, increased work productivity), although the magnitude of those benefits are difficult to calculate. It also appears that policy makers are increasingly interested in considering investments in drug abuse treatment. Much can be learned from careful evaluation of initiatives to expand treatment capacity and other interventions designed to increase participation. Unfortunately, however, this opportunity is too often overlooked.
The committee recommends increased research on the costs and benefits of policies designed to increase treatment utilization by
taking advantage of the natural experiments that are likely to occur in the years ahead. A key goal of these studies should be to assess the impact of increased treatment utilization on reducing the demand for drugs.
Also enormously informative would be purposefully planned and executed experiments or demonstration projects conducted at a community level in which treatment capacity or utilization was varied in a systematic manner. Again, the key goal would be to assess the effects of treatment expansion on reducing the demand for drugs.
Boyd, C.J., and S.E. McCabe. (2008). Coming to terms with the nonmedical use of prescription medications. Substance Abuse Treatment, Prevention, and Policy, 3.
Caulkins, J.P. (2001). Drug prices and emergency department mentions of cocaine and heroin. American Journal Public Health, 91(9), 1,446-1,448.
Caulkins, J.P., and S. Chandler. (2006). Long-run trends in incarceration of drug offenders in the U.S. Crime and Delinquency, 52(4), 619-641.
Dave, D. (2006). The effects of cocaine and heroin price on drug-related emergency department visits. Journal of Health Economics, 25(2), 311-333.
Galea, S., and M. Tracy. (2007). Participation rates in epidemiologic studies. Annals of Epidemiology, 17(9), 649-653.
Grucza, R.A., A.M. Abbacchi, T.R. Przybeck, and J.C. Gfroerer. (2007). Discrepancies in estimates of prevalence and correlates of substance use and disorders between two national surveys. Addiction, 102(4), 623-629.
Laub, J., and R. Sampson. (2003). Shared Beginnings, Divergent Lives: Delinquent Boys to Age 70. Cambridge, MA: Harvard University Press.
Johnson, L.D., P.M. O’Malley, J.E. Schulenberg, and J.G. Bachman. (2006). The Aims and Objectives of the Monitoring the Future Study and Progress Toward Fulfilling Them as of 2006. Monitoring the Future Occasional Paper Series. Available: http://monitoringthefuture. org/pubs/occpapers/occ65.pdf [accessed July 2010].
Moffitt, T.E., A. Caspi, A. Taylor, J. Kokaua, B.J. Milne, G. Polanczyk, and R. Poulton. (2010). How common are common mental disorders? Evidence that lifetime rates are doubled by prospective versus retrospective ascertainment. Psychological Medicine, 40(6), 899-909.
National Research Council. (2001). Informing America’s Policy on Illegal Drugs: What We Don’t Know Keeps Hurting Us. Committee on Data and Research for Policy on Illegal Drugs. C.F. Manski, J.V. Pepper, and C.V. Petrie (Eds.). Committee on Law and Justice and Committee on National Statistics, Commission on Behavioral and Social Sciences and Education. Washington, DC: National Academy Press.
Office of National Drug Control Policy. (2001). What America’s Users Spend on Illegal Drugs, 1988-2000. Washington, DC: The Executive Office of the President.