Our Methodology — How We Evaluate Evidence
We use a four-tier evidence framework that differentiates strong evidence from weak evidence and flags uncertainty explicitly. When we say "research suggests," we flag preliminary evidence. When we say "research demonstrates," we require moderate-to-high quality RCTs or systematic reviews. This page explains how we apply these standards.
The Framework
HIGH: Large well-designed RCTs or comprehensive systematic reviews with low bias. MODERATE: Smaller RCTs, well-designed observational studies. LOW-MODERATE: Observational with limited controls, small-sample RCTs. LOW: Case series, open-label, preclinical only. NONE: No published clinical trials. AGAINST: Evidence consistently shows harm.
Our Evidence Tiers
HIGH Evidence
Reserved for conclusions supported by:
- Large, well-designed randomized controlled trials (typically N ≥ 100 per arm)
- Comprehensive systematic reviews with multiple high-quality studies
- Meta-analyses with low risk of bias
- Replicated findings across independent research groups
Very few conclusions about cannabis meet this standard. The research base is not mature enough to support high-evidence ratings for most clinical questions.
MODERATE Evidence
Supported by:
- Smaller randomized controlled trials (N < 100 per arm)
- Well-designed observational studies with appropriate controls
- Systematic reviews with some limitations
- Mechanistic evidence supported by clinical correlation
The Bonn-Miller MAPS/Sisley trial (N=80), the Jetly 2015 nabilone trial (N=10), and the Nugent VA chronic pain review are examples of moderate-evidence sources.
LOW-MODERATE Evidence
Supported by:
- Observational studies with limited controls
- Small-sample RCTs with methodological limitations
- Cross-sectional studies showing associations
- Ecological analyses at the population level
Most cannabis evidence for specific veteran conditions falls in this tier.
LOW Evidence
Supported by:
- Case series and case reports
- Open-label pilot studies
- Retrospective chart reviews
- Animal or cell culture studies only
- Preclinical mechanism without human clinical data
Low evidence is useful for hypothesis generation but does not support clinical recommendations.
NONE Evidence
Used when:
- No published clinical trials exist for the specific question
- Research is entirely preclinical or anecdotal
- The question has not been meaningfully studied
Research gaps are honestly labeled rather than papered over with weak evidence. For example, there is essentially no research on cannabinoids for Agent Orange-related conditions — we label this gap explicitly rather than pretending that general cannabis evidence applies.
EVIDENCE AGAINST
Used when:
- Research consistently shows harm
- Research consistently fails to demonstrate benefit in rigorous trials
- The harm profile clearly outweighs any potential benefit
Cannabis use disorder, cannabis-associated suicidality, and cannabis use in people with psychotic disorder history are examples where we use "evidence against" ratings.
Language Conventions
"Research suggests"
Used when evidence is preliminary, observational, from studies with significant limitations, or otherwise low-to-moderate quality. Implies uncertainty about causation and about generalizability.
"Research demonstrates"
Used when at least moderate-quality RCT or systematic review evidence exists. Implies more confidence in causation and generalizability.
"Evidence supports"
Similar to "research demonstrates" but can apply to multiple types of evidence synthesized together.
"Contested" / "Mixed evidence"
Used when different studies or reviews reach different conclusions, indicating that the underlying question is not yet resolved.
"No direct evidence"
Used when research specifically addressing the question has not been conducted, even if related evidence exists.
Why Evidence Quality Matters
The temptation in any educational content is to overstate what we know. Evidence hierarchies exist because different study designs produce different confidence levels, and treating all evidence as equivalent produces misleading conclusions.
For cannabis specifically, the evidence hierarchy matters because:
- Decades of Schedule I restrictions produced a research base skewed toward observational studies and small trials
- The NIDA monopoly on research cannabis (1968–2021) produced material at half the potency of commercial products, limiting generalizability
- Cannabis intoxication is obvious to trial participants, making true double-blinding nearly impossible
- Self-selection biases pervade observational cannabis research
- Industry interests and advocacy organizations both have incentives to cherry-pick supportive studies
Honest evidence evaluation is the only way to provide veterans with information they can actually use to make decisions.
When We Disagree With Other Sources
Some of our conclusions differ from conclusions that appear in advocacy materials or promotional content:
- "Cannabis effectively treats PTSD" is commonly claimed; the only completed RCT of smoked cannabis in veterans found no advantage over placebo. We say this clearly, even though it complicates the advocacy narrative.
- "Cannabis prevents veteran suicide" is based on ecological studies that were reversed in later analyses. We explain this context rather than repeating the misused finding.
- "Cannabis is safer than prescription drugs" is partially true (no fatal overdoses) but obscures real harms (CUD, drug interactions, cognitive effects). We engage with the nuance.
Our goal is not to undermine cannabis advocacy or to defend VA institutional positions. It is to accurately represent what peer-reviewed research shows so that veterans can make informed decisions. Where advocacy and evidence diverge, we follow the evidence.
How We Handle Uncertainty
When evidence is uncertain, we say so. This sometimes produces content that reads less confidently than other sources. We consider this a feature, not a bug:
- A reader who assumes cannabis "works" for PTSD may make decisions they regret when it does not
- A reader who assumes cannabis is harmless may miss real risks
- A reader who understands the uncertainty can weigh it against their personal situation
Honest uncertainty is more useful than false certainty, particularly for decisions that veterans are making about their own health.
Updates and Corrections
Evidence evolves. New research is published, systematic reviews are updated, and our understanding of cannabis in veterans continues to grow. We update pages when significant new evidence is available. The "Last verified" date in the footer indicates when the page was last reviewed.
We also correct errors. If you find a factual error, please contact us. We will investigate and correct anything that is wrong.
What We Cannot Tell You
Evidence synthesis cannot answer every question:
- We cannot tell you whether cannabis will work for your specific condition — individual responses vary
- We cannot tell you whether cannabis is "right" for you — that depends on your values, priorities, and circumstances
- We cannot predict how cannabis will interact with your specific medication regimen — talk to your VA provider
- We cannot give you legal advice about your specific situation — consult an attorney
- We cannot replace your VA provider's clinical judgment