TCS Daily

Gateway to Heaven?

By Iain Murray - December 11, 2002 12:00 AM

Nevada's state ballot this year included Question 9, which called for the legalization of marijuana for both sale and use. Drug Czar John Walters visited the state twice to argue against the initiative. He was quoted in the Boston Globe as calling marijuana "an addictive gateway drug," accounting for 60 percent of Americans in drug rehab.

"It's the single biggest source of dependency of any of the illegal drugs, more than twice as important as the next most important drug, cocaine," the paper quoted Walters as saying.

The proposition fell by a wide margin, 61% to 39%. It was interesting, therefore, that, of all the studies published since that result, many of which found increased evidence that marijuana was physically or psychologically harmful, the study that got the most attention was one that cast doubt on the existence of any gateway effect. Even more interesting was how it was reported.

The gateway theory suggests that marijuana somehow affects people so that they are more likely to try hard drugs than those who have not tried marijuana. The relative risk between a marijuana user and a non-marijuana user for trying hard drugs has been shown to be as high as 85 to 1. But a new study, by Andrew Morral and a team from the RAND Corporation, published in the journal Addiction, postulated a different hypothesis to explain both this relative risk, the well-documented sequential progression in humans from marijuana to other drugs and the fact that the more frequently people use marijuana, the more likely they are to use hard drugs (the 'dose-response' relationship). The gateway theory accounts for each of these facts.

Morral and his colleagues instead postulated a model of how adolescents come to use drugs. It was based on the following assumptions: that individuals have a random predisposition or propensity to use drugs, distributed normally throughout the population; that this propensity correlates to the risk of being able to use drugs and with the probability of using them given the opportunity; and that marijuana and hard drug use are not linked.

The model produced a relative risk for marijuana users trying hard drugs substantially greater than the actual data shows, although that is probably a case of the model being very sensitive to small variations in data. It successfully predicted the ordering of drug use and also accounted for the dose-response relationship when it assumed a high correlation between high frequency marijuana use and hard drug use. Morral and his colleagues decided that their model adequately explains drug use patterns in America, and therefore suggested that the gateway theory might be an inappropriate basis for public policy decisions.

This was not quite how the study was reported, however. It was first mentioned before official publication date (and therefore in breach of an embargo, presumably) by Bill Keller in The New York Times, in the context of John Walters' "obsession" with marijuana. The Reuters story, which was also picked up by The Washington Post, got the story flat out wrong, stating that the study proved that marijuana did not lead to hard drug use and that the gateway theory was disproved. The RAND Corporation was forced to issue a press release denying this interpretation.

The gateway theory is, in fact, far from disproved. All Morral and his colleagues have done is provide an alternative explanation (it is also quite complicated, despite the authors' claims of "parsimony"). Three commentary articles accompanied the study in the journal, none of which suggested that the gateway theory was dead. One pointed out that economic evidence had shown a clear gateway effect between cigarette smoking and both alcohol and marijuana use and that research in a similar vein here could illuminate the issue. Another pointed out that even though a common underlying cause had been found that explained chickenpox, shingles and post-shingles neuralgia, nevertheless public health had been improved by targeting chickenpox first, because of the clear pathway between the three ailments.

The third commentary mentioned perhaps the most important non-statistical argument in favor of the gateway theory, the neurological data that suggest that chronic exposure to the cannabinoids in marijuana may induce cross-tolerance to harder drugs (see here for a summary of a recent conference session on the topic). Research on this topic in humans has been held up by ethical considerations, but it would fit with other neurological data about how marijuana and other drugs work. If it is a genuine effect, then it would have to be incorporated into the RAND model as a form of the gateway effect.

Although the RAND model adequately explains observed data, it cannot replace the gateway theory until further research is done. Interestingly, the Nevada quotations apart, the Drug Czar's office seeks to downplay the gateway argument, preferring instead to concentrate on the documented health effects of marijuana - marijuana does not lead to harmful drugs, it is a harmful drug. It is interesting, therefore, that of all the marijuana studies to have emerged over the past few months, this should be the one to receive major press space and that it should have been reported in such an unsatisfactory way in the nation's two most prestigious dailies.



TCS Daily Archives