Marijuana has no legitimate medical applications and is correctly classified as a dangerous Schedule I drug.

In October 1937, Harry Anslinger, the director of the Federal Bureau of Narcotics, testified before Congress in support of the Marihuana Tax Act. Among his most effective arguments was that cannabis had no legitimate medical use and that its supposed therapeutic applications were a cover for drug abuse. His testimony relied heavily on anecdote rather than pharmacology, but it achieved its goal: Congress passed the Act with little scrutiny. By then, cannabis tinctures had been falling out of clinical use in American medicine for decades, not because research had established they were ineffective, but because standardized injectables and synthetic compounds had made them less convenient. The gap between 'not currently in clinical use' and 'no medical value' was never carefully examined.

The modern framework for that claim came from the Controlled Substances Act of 1970, which placed marijuana in Schedule I, reserved for substances with high abuse potential and no accepted medical use. Schedule I status made clinical research on marijuana exceptionally difficult: researchers needed special DEA licensing, supply was controlled through a single government-authorized facility at the University of Mississippi, and institutional review boards were reluctant to approve studies involving Schedule I substances. The classification that was supposed to reflect a scientific finding also operated, in practice, to prevent the scientific investigation that would have tested it.

In health classrooms of the late 1970s and 1980s, the Schedule I designation was taught as established science. Marijuana was dangerous, addictive, and therapeutically worthless, the law said so, and the law was presented as reflecting expert consensus. What students were not taught was that the law had preceded any systematic evidence review, that the Shafer Commission appointed by Nixon had recommended against the Schedule I classification in 1972, or that a growing body of clinical observation was already complicating the official picture.

That evidence included work by Donald Abrams at the University of California San Francisco, who had been attempting since the early 1980s to conduct trials of cannabis in AIDS patients experiencing profound wasting and appetite loss. The DEA consistently blocked or delayed his research applications for years. It included observations from oncologists who noticed that patients who used marijuana before chemotherapy reported substantially less severe nausea and vomiting, findings eventually confirmed in controlled trials by researchers including Stephen Sallan at the Dana-Farber Cancer Institute, whose work in the mid-1970s showed that oral THC outperformed standard antiemetic drugs. And it included the case of Robert Randall, a glaucoma patient who successfully argued in federal court in 1976 that his use of marijuana to control intraocular pressure constituted a medical necessity, prompting the government to create a Compassionate Investigational New Drug program that it then quietly restricted to existing patients.

The political resolution came through states, not federal agencies. California's Proposition 215, the Compassionate Use Act, passed in November 1996 with 56 percent of the vote, making California the first state to authorize medical marijuana for seriously ill patients. It was driven largely by AIDS activists, cancer patients, and advocates who had spent years watching patients benefit from a substance that was classified as having no benefit. Oregon, Washington, Alaska, and Maine followed within the same election cycle.

By 1999, the Institute of Medicine's comprehensive review had identified substantial evidence that cannabis was effective in managing pain, nausea, and appetite loss, and recommended formal investigation through clinical trials. The report noted directly that the Schedule I classification was an obstacle to that investigation rather than a product of it, a federal determination made in 1970 without an evidence base sufficient to support it, and maintained for nearly three decades by the circular logic of a system that prevented the research that would have revised it.