Our research coverage begins with the question a study actually tested. We do not upgrade a laboratory result into a patient outcome or a statistical association into proof of cause.
Source order
- Peer-reviewed systematic reviews, meta-analyses, randomized trials, and major observational studies.
- Registries and primary records such as PubMed, ClinicalTrials.gov, FDA, NIH, CDC, and SAMHSA.
- Government regulatory documents, court records, and official datasets.
- University releases only as navigation to the underlying paper.
- Commercial statements only for claims about that company, never as independent proof.
Evidence labels
| Label | What it means | What it cannot establish alone |
|---|---|---|
| Systematic review | A structured synthesis of multiple studies with stated inclusion criteria. | It cannot repair weak or highly inconsistent underlying studies. |
| Randomized trial | Participants are assigned to interventions, improving causal inference. | A small or short trial may not generalize to different products, doses, or populations. |
| Observational study | Measures real-world relationships without assigning exposure. | Association does not prove cannabis caused the outcome. |
| Case report or series | Documents a clinical observation that may identify a signal. | It does not estimate typical benefit, risk, or frequency. |
| Animal or laboratory study | Tests mechanisms under controlled conditions. | It does not demonstrate a treatment works or is safe in people. |
Required extraction
For substantive research coverage we record the population, sample size, comparator, cannabis or cannabinoid formulation, dose, route, duration, outcome, uncertainty interval when reported, funding, conflicts, limitations, and whether the result matches the broader literature.
Language rules
We reserve words such as “caused,” “prevented,” and “treated” for evidence that supports those conclusions. We use “was associated with,” “participants reported,” or “the study found” when that is what the design permits. Statistical significance is not treated as proof of clinical importance.
Updates and corrections
Material new evidence can change a conclusion. Reviews should display a publication or update date and link to primary records. Corrections follow the public corrections policy.