When clinicians read the JAMA cannabis review closely, one thing becomes immediately clear. Most of the weight of its conclusions rests not on experimental certainty, but on the limits of the data it had available to summarize.
That doesn’t make the review wrong. But it does mean the conclusions need to be handled with care.
A large share of the evidence cited is observational, cross-sectional, or retrospective. Cannabis exposure is usually self-reported. Often it’s reduced to simple categories like “use” or “no use,” or “daily” versus “non-daily.” Details that matter clinically, dose, potency, route, formulation, timing, are rarely captured with much precision.
That matters because even sophisticated statistical adjustment cannot fully correct for what isn’t measured. Residual confounding remains substantial. Exposure misclassification is almost guaranteed. And as every epidemiologist knows, associations generated this way are hypothesis-generating, not proof of causality.
From a clinica…
