Note: Subscriptions to the FAS Drug Policy Analysis Bulletin may be requested by sending email to fas@fas.org, please be sure to reference the publication title
Drug Prevention: The Paradox of Timing
by Jonathan Caulkins
This essay asks whether primary prevention programs, such as school-based prevention, can play a major role in controlling drug epidemics. It concludes that in theory they could, but only if they are in place a decade or more before the epidemic is recognized. Since keeping drug prevention programs vibrant before and between epidemics may be difficult, this suggests considering whether drug prevention might best be delivered as part of a broader program intended to discourage a variety of risky behaviors.
The Good News on Prevention
Preventing an initiation into drug use is valuable. Historically, the average career of cocaine use has involved consumption of 300-400 grams of cocaine. Rydell and Everingham (1994) estimate that in 1992, $23B in social costs were associated with the 291 metric tons of cocaine consumed in the US, or about $80 per gram. That suggests that preventing an average cocaine initiation averts on the order of $24,000 - $32,000 in social costs.
Skeptics might argue that prevention programs are more likely to prevent initiation for people who would have used smaller quantities and used in fewer destructive ways than average Drug Prevention, Caulkins p.1; Broadening the Target of Prevention, Lynskey p.4users. On the other hand, in addition to completely preventing some initiations, prevention programs probably convince others who do still initiate to use less and less destructively than they otherwise would have.
So preventing drug initiation is probably valuable. Can it be accomplished? The answer seems to be yes. Although many prevention programs have no detectable effect on use, there is clear evidence from controlled studies that others do. Even interventions involving just 10-30 contact hours can affect not only attitudes and knowledge about drugs, but also use of cigarettes and marijuana (Ellickson and Bell, 1990). Furthermore, the effects of good junior high school interventions have been observed to persist at least into the 12th grade (Botvin et al., 1995).
Whether the magnitude of these effects justifies the cost of the program is difficult to determine for a variety of reasons. It is hard to predict changes in consumption of cocaine and heroin from observed changes in the use of alcohol, cigarettes, and marijuana. It is hard to predict changes in lifetime consumption from behavioral changes observed only through high school. Estimates of program costs for school-based programs depend substantially on how one values the opportunity cost of not being able to use scarce class time to teach other subjects. Despite these uncertainties, at present the best point estimate is that school-based drug prevention is cost-competitive with other interventions. Preventing drug use in the first place is more appealing than trying to address the problem later with enforcement or treatment, so it seems like primary prevention should have a prominent place in a comprehensive drug control strategy.
The Bad News for Prevention
Even if prevention can be cost-effective, for several reasons it is not clear that it can play a significant, let alone decisive, role in controlling drug use. First, prevention programs are not like vaccinations that inoculate people against infection. Even the best prevention programs do not eliminate use; significant numbers of youth who receive best practice programs go on to use drugs.
Second, rates of drug use over time display an epidemic character, with phases of incubation, expansion, plateau, and decline (Golub and Johnson, 1997). One can not prevent that which has already occurred, so prevention programs late in this cycle, when initiation rates have already receded, cannot address the majority of the problem. Everingham and Rydell (1994) estimated that even if no one initiated cocaine use after 1992, US annual consumption would still be about 150 metric tons fifteen years later. Their model suggests that eliminating initiation would sharply reduce the number of light users, but it would have a much more modest effect on heavy users because heavy users persist in their use for many years. Since heavy users are responsible for the majority of the cocaine use and related problems, this places an inherent limit on what prevention can contribute in the later phases of a drug epidemic.
Furthermore, early in an epidemic use spreads like a contagion with existing users introducing others to the drug. At that time, convincing one person to abstain can preempt a cascade of initiations. Later, this chain reaction is much less explosive because there is greater awareness of the risks and problems of drug use, so users "recruit" initiates much more slowly. The magnitude of the difference in effectiveness between prevention early and late in an epidemic depends on one's particular vision of the initiation process. However, for any model in which "light" users promote initiation, "heavy" users discourage initiation, and light users become heavy users over time, prevention will be more effective early in an epidemic. For the models I have examined, the difference ranges from substantial to huge. The models also suggest that good prevention programs that affect early stages of an epidemic can significantly reduce the intensity and duration of that epidemic.
Preventing initiations early in an epidemic is desirable, but not necessarily feasible. Our "advance warning" and "monitoring" systems have a hard time detecting an epidemic in its early stages, and there is next to no capacity to predict the arrival of an epidemic ahead of time. Public awareness, concern, and willingness to act are driven by problems associated with use, and those problems peak after initiation does. For example, cocaine initiation peaked in the late 1970s, but cocaine was not widely recognized as a major problem until the early to mid 1980s.
There are further delays before a program designed to respond to that problem can be in the field. Funds must be budgeted, appropriated, and disseminated to line agencies. Model programs must be developed and expanded to national scale. Combined, these delays would probably take at least three years.
The greatest delay is between the implementation of a prevention program and when its effects begin to be felt. School-based prevention programs often target 7th or 8th graders, who are typically 13 or 14 years old. The median age of cocaine initiation is 21.5 years (Johnson et al., 1996), so there is roughly an eight year delay between when a prevention program is implemented and when its effects on cocaine initiation begin to be felt. Since most heavy users go through a period of light use before progressing to heavy use, the delay between implementation and the first significant effects on use are even greater.
The Policy Problem
If the cocaine epidemic was a one-time event, it is too late for primary drug prevention to make much of a difference. The horse is already out of the barn. However, drug epidemics come in cycles with both long periods (Musto (1987) compares the late 19th and late 20th century drug epidemics) and shorter periods (30-day prevalence of marijuana use among high school seniors peaked in 1979, fell to a trough in 1992, and has risen sharply since). So even if prevention can not do much to help with the current cocaine epidemic, it could play a significant role in future epidemics. The value of prevention will be especially great if there are future epidemics with high initiation rates and in which initiation by one person creates a cascade of other initiations.
We could wait to deploy prevention programs until there is evidence that such an epidemic has started. But if the next epidemic behaves like the last one, this relegates primary prevention to a fairly minor role and forfeits its greatest potential for being cost-effective.
Alternately, we could put prevention programs in place now and maintain them indefinitely, in the hope they will help stave off an epidemic in 20 or 40 years. With that approach, prevention could make a big difference, but large scale prevention programs would be funded for many years when initiation rates would have been low anyhow. Furthermore, inasmuch as the prevention program succeeds in preventing future epidemics, the public may grow impatient with spending so much when there is no apparent drug menace. On the other hand, if there were still a subsequent drug problem, even if it were smaller than it otherwise would have been, the program would be vulnerable to criticisms that it did not work; counterfactuals are hard to establish in such circumstances.
Proposed Solution
One possible resolution to this conundrum would be to run prevention programs at all times, but not to focus them on drugs or call them "drug" prevention programs. Convincing kids not to use drugs has a lot in common with convincing them to resist other temptations and to consider long term consequences in other contexts. Crime prevention, gang prevention, drop out prevention, preventing running away from home, discouraging premarital sex (particularly unprotected sex and teen-age child bearing), and discouraging drinking and driving are all similar endeavors. Drug prevention also has a lot in common with positive interventions designed to get people to invest in their own well being, including health promotion, encouraging exercise, encouraging good nutrition, promoting good school performance, etc. Indeed, many existing prevention programs-- including some school-based programs but even more some community-based and alternatives-based programs--try to generate such diverse benefits already (Caulkins et al., 1994).
There are a number of advantages to such an approach. First, drug epidemics come and go, but youth struggling with some sort of temptation is a perennial theme in human history. The more comprehensive the prevention message, the greater the likelihood that any given student will benefit in some way. Second, for some kinds of programs the marginal cost of adding an additional prevention message (e.g., gang prevention) to an existing program that delivers some other prevention message (e.g., drug prevention) is much less than the cost of starting a new program dedicated to the new message. Hence, it could be more efficient to bundle the various prevention messages together in one program than it would be to have every child go through a half dozen or more separate programs.
Third, even if a prevention program is designed with one purpose in mind, it will often generate other benefits as well. E.g., a drug prevention program may improve high school completion rates, and a drop out prevention program might reduce drug use. If a prevention program is promoted under just one banner, then society may overlook or undervalue these collateral benefits and, as a result, tend to under-invest in the program.
Fourth, it might even turn out that combining programs would make them better, although that is an empirical question. Certainly research efforts could be combined. The broader agenda might also focus attention on core values (temptation, time horizons, peer pressure, etc.) and at the same time be less moralistic. Arguing that kids should abstain from drugs may be a more persuasive message when delivered as part of a general program that encourages healthy behaviors than when it is given in isolation.
It may be time to extend DARE into TREK and invest in Temptation Resistance Education for Kids, not just Drug Abuse Resistance Education.
References:
Botvin, Gilbert J., E. Baker, L. Dusenbury, E.M. Botvin, and T. Diaz [1995], "Long-Term Follow-Up Results of a Randomized Drug Abuse Prevention Trial in a White Middle-Class Population," Journal of the American Medical Association, Vol. 273, pp. 1106-1112.
Caulkins, Jonathan P., Nora Fitzgerald, Karyn Model and H. Lamar Willis [1994], "Youth Drug Prevention Through Community Outreach: The Military's Pilot Programs," MR-536-OSD, RAND, Santa Monica, CA.
Ellickson, Phyllis L. and Robert M. Bell [1990], "Drug Prevention in Junior High: A Multi-Site Longitudinal Test," Science, Vol. 247 (March), pp. 1299-1305.
Golub, Andrew Lang and Bruce D. Johnson [1997], "Crack's Decline: Some Surprises Across U.S. Cities," National Institute of Justice Report NCJ 164262, Washington, DC.
Musto David F. [1987], The American Disease: Origins of Narcotic Control. New York: Oxford University Press.
Broadening the Target of Drug Prevention
by Michael T Lynskey
There is continued public support for school-based programs that aim to delay or prevent the onset of substance use during adolescence and young adulthood. These programs have been motivated by the widespread use of tobacco, alcohol and illicit drugs, and by the resulting harm. Not only does hazardous use during adolescence puts young people at risk of experiencing negative short-term consequences of substance use, but the early onset of substance use is an independent risk factor for substance-related harm and problems of adjustment in later life (Lynskey & Hall, In press; Yamaguchi & Kandel, 1984).
Unfortunately, public enthusiasm for these programs is not matched by evidence of their efficacy. Investigations of the effectiveness of a number of these programs have concluded that many do not influence substance use; in some cases, programs actually seem to increase experimentation with tobacco, alcohol and illicit drugs. However, some more recent studies have established that school-based prevention programs can delay or avoid the onset of substance use (Botvin, Baker, Dusenbury, Botvin & Diaz, 1995; Ellickson, Bell & McGuigan, 1993). Among the factors which are likely to optimize their effectiveness are: "booster" sessions conducted over a number of years (Botvin et al, 1995; Ellickson et al, 1993); an emphasis on social skills training and general life skills rather than on drug knowledge (Botvin et al, 1995; Ellickson et al, 1993; Tobler, 1986); and the active inclusion of peers in running the programs (Tobler, 1986). More recently, a meta-analysis has shown the superiority of programs which are delivered interactively and which focus on interpersonal competence and peer influences (Tobler, 1997).
Existing programs tend to focus on preventing or delaying the onset of substance use. This concentration is reflected in the programs' design goals, in the messages they actually deliver to their young audiences, and in the evaluation criteria by which their success is measured.
This narrow focus may be a mistake. Many studies have demonstrated that a wide range of health-compromising "problem" behaviors cluster or co-occur. Young people who report early substance use behaviors are at increased risk for precocious sexual activity and sexual risk-taking, for criminal offending, and for mental health problems. The origins of the associations between a number of pairs of behaviors or disorders have been examined in a series of studies conducted as part of the Christchurch Health and Development Study. The behaviors studied have included tobacco use and depressive symptomatology (Fergusson, Lynskey & Horwood, 1996a); alcohol use and criminal offending (Fergusson, Lynskey & Horwood, 1996b) and alcohol use and sexual risk-taking (Fergusson & Lynskey, 1996). These studies have confirmed that there are moderate to high degrees of association between each pair of behaviors. Further analyses indicated that in each case a large component of the associations between these behaviors could be explained by the influence of shared or common risk factors, observable throughout childhood, which increase the risks of each outcome. The risk factors for each of these outcomes have been identified previously and include affiliation with delinquent or substance-using peers; family substance use behaviors and exposure to family dysfunction; individual temperamental, personality and genetic factors; and social disadvantage and exposure to adverse family living conditions.
These findings have several implications for the design and implementation of prevention programs. Firstly, such programs should address, both in the messages they deliver and in the aims they pursue, a wider range of potentially harmful behaviors such as sexual risk-taking and criminal offending. Broadening the topics addressed has the benefit of giving children a better picture of the full range of health-risk behaviors with which they are likely to be tempted. Not only is a reduction in these behaviors itself a desirable goal, but the reduction or amelioration of these behaviors may, in turn, lead to a reduction in substance use and substance-related harm.
Recognizing that much of the association between different problem behaviors arises from the influence of common risk factors, interventions should target and attempt to modify the known risk factors for substance use and other behavioral disturbances. One such study has been reported by Tremblay et al (1995) who implemented a comprehensive intervention program for at-risk kindergarten aged boys that included both parent training and social skills training for the boys. At follow-up in adolescence the intervention group showed significantly reduced rates of aggressive behavior and increased grade retention. While substance use was not assessed at follow-up, it is reasonable to hope that that which successfully targets known risk factors would lead to a reduction in later substance misuse. This is an attractive research topic. Conversely, since successful substance abuse prevention programs are typically characterized by an emphasis on social skills training and general life skills, it is reasonable to suggest that these programs may also have positive benefits in reducing the incidence of other health compromising behaviors; again, it would be useful to test this speculation empirically. The extent to which existing drug-prevention programs may have beneficial effects in other areas remains largely undetermined, because evaluations of these programs have typically focused solely on their effects on drug use. Thus, as the targets of school-based prevention programs are broadened to include consideration of non-drug-related behaviors, evaluations of these programs should also be expanded to measure their impacts on crime, sexual risk-taking, and other risky activities.
Finally, there is a need to develop special interventions for high-risk groups, as those most at risk are often the least likely to receive broad-based programs. For example, previously in New Zealand a comprehensive range of publicly-funded social programs ensured population-wide access to free child health services and universal access to preschool education. The basic aim of these services was to insure that all children received equal access to high quality health and educational services, regardless of their social background. However, the Christchurch Health and Development Study found that families most in need of additional assistance were those least likely to access such services. This confirmation of the "law of inverse care" suggests that population-based initiatives should be designed to ensure that those most at risk for substance use problems are not systematically excluded from participation in these programs.
In summary, while we should intervene to prevent or delay the onset of substance use behaviors during adolescence, such interventions should not focus solely on substance use. They should target a range of potentially health threatening behaviors including substance use, sexual risk-taking and problems of personal adjustment since many of these behaviors co-occur because of shared risk factors. Given that the small percentage of the population who has high risks of these adverse outcomes is least likely to access traditional public health-based interventions, population-based interventions should be supplemented by efforts to maximize access to these programs by those most at risk.
References:
Botvin, G. J., Baker, E., Dusenbury, L., Botvin, E. M., & Diaz, T. (1995). Long-term follow-up results of a randomized drug abuse prevention trial in a white middle-class population. Journal of the American Medical Association, 273, 1106-1112.
Ellickson, P. L., Bell, R. M., & McGuigan, K. (1993). Preventing adolescent drug use: Long-term results of a junior high program. American Journal of Public Health, 83, 856-861.
Fergusson, D.M. & Lynskey, M.T. (1996). Alcohol misuse and adolescent sexual behaviors and risk taking. Pediatrics, 98, 91-96.
Fergusson, D.M., Lynskey, M.T. & Horwood, L.J. (1996a). The comorbidity between affective disorders and nicotine dependence in a cohort of 16 year olds. Archives of General Psychiatry, 53, 1043-1047.
Fergusson, D.M., Lynskey, M.T. & Horwood, L.J. (1996b). Alcohol misuse and juvenile offending in adolescence. Addiction, 91, 483-494.
Lynskey, M. T., & Hall, W. (In press). Cohort trends in age of initiation to heroin use. Drug and Alcohol Review.
Tobler, N. S. (1986). Meta-analysis of 143 adolescent drug prevention programs: Quantitative outcome results of program participants compared to a control or comparison group. Journal of Drug Issues, 16, 537-567.
Tobler, N. S. (1997). Meta-analysis of adolescent drug prevention programs: Results of the 1993 meta-analysis. In W. J. Bukoski (Ed.), Meta-analysis of drug abuse prevention programs. NIDA Research Monograph 170. Rockville, MD: National Institute on Drug Abuse.
Tremblay, R. E., Pagani-Kurtz, L., Masse, L. C. Vitaro, F. & Pihl, R. O. (1995). A bimodal preventive intervention for disruptive kindergarten boys: its impact through mid-adolescence. Journal of Consulting & Clinical Psychology, 63, 560-568.
Yamaguchi, K., & Kandel, D. B. (1984). Patterns of drug use from adolescence to young adulthood. III. Predictors of progression. American Journal of Public Health, 74, 673-681.
The Bulletin is edited by Mark A.R. Kleiman of the School of Public Policy and Social Research at UCLA. Please direct comments or questions about the contents of this newsletter to Dr. Kleiman at (310)206-3234 or by e-mail at kleiman@ucla.edu. Members of the editorial board include (affiliations for identification only):
Drug Policy Analysis Bulletin c/o Federation of American Scientists 307 Massachusetts Ave., NE Washington, DC 20002