The standards have been tightened since the present system was established in 1962, and few applications that were approved in the early '60s would be approved today on the basis of the same evidence. Certainly we need more laboratory and clinical research to improve our understanding of medicinal cannabis. We need to know how many patients and which patients with each symptom or syndrome are likely to find cannabis more effective than existing drugs. We also need to know more about its effects on the immune system in immunologically impaired patients, its interactions with other medicines, and its possible uses for children.

But I have come to doubt whether the FDA rules should apply to cannabis. There is no question about its safety. It is one of humanity’s oldest medicines, used for thousands of years by millions of people with very little evidence of significant toxic effects. More is known about its adverse effects than about those of most prescription drugs. The government of the United States has conducted through its National Institute of Drug Abuse (NIDA) a decades-long multimillion-dollar research program in a futile attempt to demonstrate significant toxic effects that would justify the prohibition of cannabis as a non-medical drug. Should time and resources be wasted to demonstrate for the FDA what is already so obvious?

But even if it were legally and practically possible to do the various phased studies to win FDA approval, where would the money to finance these studies come from? New medicines are almost invariably introduced by drug companies that spend many millions of dollars on the development of each product. They are willing to undertake these costs only because of the anticipated large profits during the 20 years they own the patent. Obviously pharmaceutical companies cannot patent marijuana. In fact they are very much opposed to its acceptance as a medicine because it will compete with their own products.

It is unlikely that whole smoked marijuana should or will ever be developed as an officially recognized medicine via this route. Thousands of years of use have demonstrated its medical value; the extensive government-supported effort of the last three decades to establish a sufficient level of toxicity to support the harsh prohibition has instead provided a record of safety that is more compelling than that of most approved medicines. The modern FDA protocol is not necessary to establish a risk-benefit estimate for a drug with such a history. To impose this protocol on cannabis would be like making the same demand of aspirin, which was accepted as a medicine more than 60 years before the advent of the double-blind controlled study. Many years of experience have shown us that aspirin has many uses and limited toxicity, yet today it could not be marshaled through the FDA approval process. The patent has long since expired, and with it the incentive to underwrite the substantial cost of this modern seal of approval. Cannabis, too, is unpatentable, so the only sources of funding for a "start-from-scratch" approval would be non-profit organizations or the government, which is, to put it mildly, unlikely to be helpful. Other reasons for doubting that marijuana would ever be officially approved are today’s anti-smoking climate and, most important, the widespread use of cannabis for purposes disapproved by the government.

To see some of the obstacles to this approach to the problem, consider the effects of granting marijuana legitimacy as a medicine while prohibiting it for any other use. How would the appropriate "labeled" uses be determined and how would "off-label" uses be monitored? Let us suppose that studies satisfactory to the FDA are somehow completed affirming that marijuana is safe and effective as a treatment for the AIDS wasting syndrome and/or AIDS-related neuropathy, and physicians are able to prescribe it for those conditions. This will present unique problems. When a drug is approved for one medical purpose, physicians are generally free to write off-label prescriptions - that is, prescribe it for other conditions as well. If marijuana is approved as a medicine, how will off-label prescribing play out? Surely, knowledgeable physicians will want to prescribe it for some patients with multiple sclerosis, Crohn's disease, migraine, convulsive disorders, spastic symptoms, and other conditions for which the use of cannabis is well established by a mountain of anecdotal evidence. But what about premenstrual syndrome? Surely women who suffer from this disorder consider it a serious problem, and many of them find cannabis the most useful and least toxic treatment. What about the loss of erectile capacity in paraplegics? What about intractable hiccups? And then there is depression, not the DSM-IV defined major affective disorder, but the common low-level dysphoric condition for which general practitioners frequently prescribe SSRI's such as Prozac? What about bipolar disorder?

Generally speaking, the more dangerous a drug is, the more serious or debilitating must be a symptom or illness for which it is approved. Conversely, the more serious the health problem, the more risk is tolerated. If the benefit is very large and the risk very small, the medicine is distributed over the counter (OTC). OTC drugs are considered so useful and safe that patients are allowed to use their own judgment without a doctor's permission or advice. Thus, today anyone can buy and use aspirin for any purpose at all. This is permissible because aspirin is considered to be so safe; it takes "only" one to two thousand lives a year in the United States. The remarkably versatile ibuprofen (Advil) and other non-steroidal anti-inflammatory drugs (NSAIDs) can also be purchased OTC because they, too, are considered very safe; "only" 10,000 Americans lose their lives to these drugs annually. Acetaminophen (Tylenol), another useful OTC drug, is responsible for about 10 percent of cases of end-stage renal disease. The public is also allowed to purchase many herbal remedies whose dangers and efficacies have not been well determined. Compare these drugs with marijuana. Today, no one can doubt that it is, as DEA Administrative Judge Francis L. Young put it,"... among the safest therapeutic substances known to man." If it were now in the official pharmacopoeia, it would be a serious contender for the title of least toxic substance in that compendium. In its long history, cannabis has never caused a single overdose death.