This is the message that emerges from a new RAND study that offers the first probing analysis of the role that drug prevention education can be expected to play in mitigating the drug crisis. The study appears amidst increasing doubts that the nation can arrest and incarcerate its way out of the problem and as opinion polls reflect growing support for prevention. The federal government recently launched a $195 million a year media campaign "to educate and enable our youth to reject substance abuse." But how effective is prevention and at what price relative to other drug control strategies?

Because data are lacking on the effects of media and community-based prevention strategies, the research team focused on several successful school-based prevention experiments with documented short-term effects on marijuana, tobacco and alcohol use. They estimate that these model programs not only reduce consumption of cocaine—the most problematic illicit drug—but can do so just as cost-effectively as some law enforcement efforts do. Moreover, the strategy is plainly affordable—a nationwide push to cover all youth would cost $550 million annually, or just 1.5 percent of national drug control spending.

But the researchers also found that the effects of implementing such a program would be modest and slow to accrue. It would reduce the lifetime cocaine consumption of program participants by between 2 percent and 11 percent. At that rate, it would take 10 years to reduce the number of cocaine users by 2.5 percent, 20 years to reduce the number by 5 percent and 40 years to achieve a 7.5 percent reduction.

One implication: Proposals to legalize cocaine and prevent any increase in use by shifting resources from enforcement to prevention are unrealistic.

The title of the study—An Ounce of Prevention, a Pound of Uncertainty—is a pointed one. A key finding is that precise estimates cannot be provided given the limitations of past research and current data. (This study is the first, for example, to estimate prevention's effects on lifetime drug use and on the associates of program participants.) Because of the data limitations, the researchers designed their analysis to be modular; as better information becomes available, it can be plugged in to provide updated estimates with narrower uncertainty bands.

At present, the best estimate regarding cost-effectiveness suggests that prevention can cut more cocaine use per million dollars of spending than some control programs—notably crop eradication, interdiction and long prison sentences for typical drug dealers—while falling short of others. But the range of reasonable estimates is wide enough that these comparisons cannot be made with a great deal of confidence.

Prevention might have accomplished more had it been undertaken before the current cocaine epidemic got under way. Prevention programs are most effective at the onset of a drug epidemic when initiation is high and most contagious. (The contagious spread of drug use implies that preventing one initiation can short circuit a chain reaction of initiations.) Today the epidemic that began in the late 1960s is mature.

"At this stage, prevention can't play a decisive role," summarizes lead author Jonathan P. Caulkins. "It isn't the silver bullet everyone would like it to be. Still, school-based programs can play a worthwhile supportive role with respect to the current situation while providing a form of 'cheap insurance' against any future epidemics."

Underscoring that role is a finding that each dollar spent on such a program today would avert an eventual $2.40 in costs of cocaine-related crime, health care and lost productivity. Again, this is a best estimate within a broad range of uncertain outcomes. But it is highly likely that the social benefits exceed the program costs, the researchers contend, particularly if non-cocaine-related benefits are considered. These include reductions in cigarette and alcohol consumption and potential improvements in participants' general social behavior.

The study is based on past evaluations of two programs—Project ALERT (developed at RAND and currently in operation throughout South Dakota) and the Life Skills Training Program. Both motivate and train adolescents to resist peer pressures to use drugs and both have proven records of discouraging marijuana use. The researchers combined those records with data on marijuana-cocaine relationships to project future reductions in cocaine use. Then they accounted for technical factors that would decrease or increase those estimates.

To determine cost-effectiveness, the team divided the findings on reduction in cocaine use by the $150 per pupil cost of prevention classes and compared the results to previous RAND cost-effectiveness estimates for a variety of other drug control programs. Most of the cost of running the school-based programs is not direct budgetary cost but rather the opportunity cost of using school time to teach non-academic material.

The researchers emphasize that evaluating prevention solely in terms of the number of program participants who refrain from initiating drug use sells prevention short, recognizing only about one-quarter of the drug-related benefits of prevention. Other benefits come from reduced and delayed drug use by participants, from spillover effects on friends and associates, and through effects on the cocaine market.

The study was funded by the Substance Abuse Policy Research Program of The Robert Wood Johnson Foundation and conducted within RAND's Drug Policy Research Center.

##