Last verified: April 2026

The 2017 Report — Why It Still Matters

In January 2017, the National Academies of Sciences, Engineering, and Medicine (NASEM) published The Health Effects of Cannabis and Cannabinoids: The Current State of Evidence and Recommendations for Research. A committee of 16 experts chaired by Marie McCormick of the Harvard T.H. Chan School of Public Health reviewed over 10,700 abstracts, producing nearly 100 evidence-based conclusions across therapeutic benefits, risks, and research barriers.

The report was commissioned precisely because the gap between clinical evidence and state medical cannabis programs had become untenable. By 2017, 28 states had legalized medical cannabis, most listing qualifying conditions that had never been tested in rigorous clinical trials. NASEM’s task was to impose scientific order on a landscape shaped primarily by political lobbying and patient anecdote. As of April 2026, no updated NASEM review has been published, meaning the 2017 hierarchy remains the most authoritative federal assessment despite being nearly a decade old.

The Five-Tier Evidence Hierarchy

NASEM organized its therapeutic conclusions into five tiers of evidence weight. Understanding these tiers is essential for interpreting every medical cannabis claim:

Conclusive or Substantial Evidence — the strongest tier — was assigned to only three therapeutic areas: chronic pain (particularly neuropathic), chemotherapy-induced nausea and vomiting, and multiple sclerosis spasticity (patient-reported). This tier requires consistent findings across well-designed randomized controlled trials with sufficient sample sizes and replication.

Moderate Evidence was applied to conditions where data suggested benefit but methodological limitations prevented a stronger conclusion. This tier included sleep disturbance associated with conditions like chronic pain, fibromyalgia, and obstructive sleep apnea — but notably, not primary insomnia in otherwise healthy individuals.

Limited Evidence covered conditions with some supporting data but inadequate trial design, small samples, or inconsistent results. Conditions in this tier included appetite and weight gain in HIV/AIDS, Tourette syndrome tics, anxiety (specifically social anxiety disorder), and PTSD symptoms.

Insufficient Evidence was the designation for the majority of conditions, including cancer treatment, epilepsy (which has since been elevated by the Epidiolex trials), irritable bowel syndrome, ALS, Huntington’s disease, Parkinson’s disease, dystonia, and addiction to other substances. Insufficient does not mean “doesn’t work” — it means the evidence was too sparse or too poorly designed to draw any conclusion.

No or Insufficient Evidence to Support or Refute was applied to conditions where essentially no relevant data existed.

The 8.3% Problem — State Programs vs. Evidence

A 2024 analysis of qualifying conditions across all U.S. state medical cannabis programs revealed a striking disconnect: only 8.3% of state qualifying conditions meet the NASEM “conclusive or substantial” evidence standard. The vast majority of conditions for which states authorize medical cannabis use — including PTSD, anxiety, autism, Parkinson’s disease, and chronic pain categories broader than neuropathic — are rated as having limited or insufficient evidence.

This gap is not accidental. State medical cannabis programs were created through legislative and ballot initiative processes that responded to patient demand, not clinical evidence review. Qualifying conditions were typically added by medical advisory boards under political pressure, with the bar for inclusion set far below the standard required for FDA drug approval. The result is a system in which patients receive legal access to a therapeutic agent for conditions where the evidence of efficacy may be marginal, absent, or even contradicted by available data.

This is not an argument against medical cannabis access — it is an argument for honest framing. Patients deserve to know whether their condition falls in the “strong evidence” category (chronic neuropathic pain) or the “we genuinely don’t know” category (fibromyalgia), because that distinction should inform expectations, treatment planning, and the decision to continue or discontinue use.

What the Report Did Not Do

Three critical limitations of the NASEM report are frequently overlooked. First, the committee did not make treatment recommendations. It assessed evidence quality, not clinical practice. Saying that chronic pain has “substantial evidence” of benefit is not the same as recommending cannabis for chronic pain — it means the data consistently show an effect, not that the effect is large enough or safe enough to prefer over alternatives.

Second, the report assessed cannabis broadly, not specific cannabinoids, doses, routes, or formulations. The “substantial evidence for chronic pain” conclusion is based on studies using a heterogeneous mix of whole-plant cannabis, synthetic THC (dronabinol), THC:CBD sprays (nabiximols), and smoked flower. Whether a particular dispensary product would produce the same result is an entirely separate question the report was not designed to answer.

Third, the committee explicitly called for Schedule I reform to enable the very research that would resolve the evidence gaps. The catch-22 is self-evident: cannabis remains in the category reserved for substances with “no accepted medical use,” yet federal restrictions on research prevent the clinical trials that could establish accepted medical use. The committee noted that this regulatory barrier is the single largest obstacle to building a stronger evidence base.

Evidence Evolution Since 2017

The most significant evidence development since the NASEM report is the Epidiolex approval (June 2018), which moved epilepsy from “insufficient evidence” to what would now be considered conclusive evidence — specifically for CBD in Dravet syndrome, Lennox-Gastaut syndrome, and tuberous sclerosis complex. This is the only condition category that has clearly moved up the evidence hierarchy since the report’s publication.

The opioid-sparing hypothesis has accumulated additional observational data since 2017, with ecological studies showing associations between medical cannabis access and reduced opioid prescribing and overdose mortality. However, the data remain observational and ecological — not the randomized trial evidence that would move the needle on the NASEM hierarchy.

The PTSD evidence has, if anything, weakened since 2017. The most rigorous RCT to date — Bonn-Miller et al. (2021) — found no significant difference between any cannabis group and placebo in 80 veterans, despite PTSD being a qualifying condition in the majority of medical cannabis states.

Cancer treatment remains at “insufficient,” though the ARISTOCRAT Phase 2 trial (nabiximols plus temozolomide for recurrent glioblastoma) is ongoing and may produce data sufficient to upgrade the classification for that specific indication. No broader anti-cancer evidence has emerged from clinical trials.

Using This Section

The following pages in this Medical Evidence section examine each major therapeutic area in detail: the specific trials, named researchers, effect sizes, methodological limitations, and honest assessment of where the science actually stands. Each page identifies what the evidence does support, what it does not, and what questions remain unanswered. The goal is not to advocate for or against medical cannabis, but to provide the level of scientific detail necessary for informed decision-making — something the current system of dispensary consultations and internet claims fails to deliver.