Evidence Quality Framework — How to Read Our Ratings

Our four-tier evidence rating system appears throughout this site. This page explains each tier in detail so you can interpret our ratings correctly and understand what we mean when we label evidence as HIGH, MODERATE, LOW-MODERATE, LOW, NONE, or AGAINST.

Quick Reference

HIGH: trust it. MODERATE: take it seriously. LOW-MODERATE: interesting but not decisive. LOW: hypothesis-generating only. NONE: honestly, we do not know. AGAINST: evidence points toward harm or failure to show benefit.

Why Evidence Hierarchies Exist

Different research designs produce different levels of confidence. A randomized controlled trial with 1,000 participants and proper blinding produces far more reliable conclusions than a retrospective chart review of 20 patients. Lumping all "research" together treats these as equivalent, which they are not.

Evidence hierarchies are standard tools in evidence-based medicine. The specific tiers we use are adapted to the unique limitations of cannabis research, which has been skewed for decades by federal Schedule I restrictions.

The Tiers in Detail

HIGH Evidence ⭐⭐⭐⭐

Definition: Multiple large, well-designed randomized controlled trials, or comprehensive systematic reviews with low risk of bias, or meta-analyses that synthesize high-quality evidence.

Confidence level: We have strong confidence in the conclusion. Future research might refine it, but major reversals are unlikely.

Example: For conditions like chemotherapy-induced nausea, where pharmaceutical cannabinoids (dronabinol, nabilone) have extensive RCT evidence, we could use HIGH evidence ratings. For most veteran cannabis questions, HIGH evidence does not yet exist.

MODERATE Evidence ⭐⭐⭐

Definition: Smaller randomized controlled trials, well-designed observational studies with appropriate controls, or systematic reviews with some limitations but generally reliable conclusions.

Confidence level: We have reasonable confidence. The conclusion is likely correct, but caveats apply.

Example: The Jetly 2015 nabilone RCT in 10 Canadian Forces patients (small sample) and the Nugent 2017 VA chronic pain review (low-strength conclusion) are moderate-evidence sources. The Neumeister 2013 PET imaging study establishing endocannabinoid alterations in PTSD is another.

LOW-MODERATE Evidence ⭐⭐

Definition: Observational studies with limited controls, small-sample RCTs with methodological limitations, cross-sectional studies showing associations, or ecological analyses.

Confidence level: Suggestive but not decisive. Multiple studies may support a conclusion, but confounding and bias concerns limit confidence.

Example: Much of the observational cannabis research in veterans (Loflin, Earleywine, Bonn-Miller and others) falls in this tier. These studies provide useful information about patterns of use and self-reported outcomes but cannot establish causation.

LOW Evidence ⭐

Definition: Case series, case reports, open-label pilot studies, retrospective chart reviews with significant bias, or preclinical data only.

Confidence level: Hypothesis-generating. Should not be treated as clinical evidence but can suggest directions for further research.

Example: The Greer 2014 retrospective chart review of New Mexico medical cannabis patients (pre-selected for response) is low evidence. Preclinical endocannabinoid research without human correlates is low evidence for clinical use.

NONE

Definition: No published clinical trials or systematic reviews address the specific question.

Confidence level: We honestly do not know. Absence of evidence is not evidence of absence — but it also is not evidence of presence.

Example: Cannabinoids for Agent Orange-related conditions. Cannabinoids for military sexual trauma. These are research gaps, not conclusions.

EVIDENCE AGAINST

Definition: Research consistently shows harm, failure to demonstrate benefit in rigorous trials, or risk profiles that clearly outweigh potential benefits.

Confidence level: We are confident that the claim under discussion is not supported or is actively contradicted by evidence.

Example: Cannabis as a PTSD treatment in the only completed veteran RCT (Bonn-Miller 2021), cannabis use in the context of existing psychotic disorder, and cannabis use while holding federal security clearances are examples where we use "evidence against" framings.

Applying the Tiers to Specific Questions

The same general condition can have different evidence ratings for different specific questions. For example:

Cannabis and PTSD

Plant cannabis for PTSD: LOW-MODERATE (only RCT negative; observational studies limited)
Nabilone for PTSD nightmares: MODERATE (Jetly 2015 RCT, small but positive)
CBD for PTSD: LOW (small pilot studies, no rigorous RCT)
Cannabis as first-line PTSD treatment: AGAINST (VA/DoD CPG strongly recommends against)
Endocannabinoid system involvement in PTSD: MODERATE (Neumeister 2013 PET evidence)
Extinction learning effects of cannabis: LOW-MODERATE (mixed findings, concerning signal)

Cannabis and Chronic Pain

Cannabinoids for neuropathic pain: MODERATE (Nugent 2017, Whiting 2015, Canadian Pain Society)
Cannabinoids for musculoskeletal pain: LOW (limited direct evidence)
Cannabinoids for phantom limb pain: LOW (case reports only)
Cannabis reduces opioid use at population level: MIXED (Bachhuber positive, Shover reversed)
Cannabis as substitute for opioids in individuals: LOW-MODERATE (observational only)

Disaggregating evidence by specific question is more useful than giving single ratings to broad condition categories.

How We Handle Evolving Evidence

Cannabis research is moving. The MJP2 Phase 2 trial will produce new data in 2026–2027. Ongoing VA research continues. Trump's EO 14370 may accelerate research access. Rescheduling, if it occurs, could transform the research landscape.

We update our ratings as new evidence becomes available. This means our current conclusions could change. If we tell you today that evidence for a specific claim is LOW, and a large rigorous RCT produces HIGH-quality evidence in 2027, we will update the page. The "Last verified" date in the footer tells you when we last reviewed the page.

Why We Do Not Over-Claim

The temptation in cannabis education is to overstate what we know. Advocacy is more compelling when "research shows" rather than "research suggests." Articles are more shareable when uncertainty is eliminated. Product marketing depends on confident claims.

But over-claiming has costs:

Veterans who act on inflated evidence may be disappointed, harmed, or misled
Over-claimed benefits that fail in rigorous trials damage credibility for all cannabis research
Veterans who understand that the evidence is uncertain can make better decisions for their specific situations
Honest uncertainty respects veteran autonomy more than false certainty

We prefer to tell you the truth about what we know and do not know, even when the truth is uncomfortable.

How to Use This Framework

HIGH or MODERATE evidence supports clinical decision-making. You can rely on it in evaluating your options.
LOW-MODERATE evidence is worth considering but not decisive. It suggests rather than proves.
LOW evidence is hypothesis-generating. It tells you something has been studied but not enough to be confident about it.
NONE means you cannot make an evidence-based decision — you are operating on other considerations.
AGAINST evidence should be weighted heavily against a given course of action.