Real-world approaches are superior to RCTs for studying the potential of CBMPs to treat anxiety
- Michael T Lynskey, Larry Phillips, Anne K Schlag and David J Nutt
- 2 hours ago
- 1 min read
Authors
Michael T Lynskey, Larry Phillips, Anne K Schlag and David J Nutt
Published
July 9, 2026
Abstract
RCTs of cannabis-based medicinal products (CBMPs) for anxiety consistently show little evidence of efficacy, yet patient use continues to grow. This editorial argues the discrepancy is largely artefactual, reflecting flawed trial methodology rather than a genuine lack of benefit.
Using data from Project Twenty21, a UK observational registry, we compare outcomes for 176 patients treated for generalised anxiety disorder against a recent meta-analysis of four anxiety RCTs (n=292). Real-world patients showed substantial improvements in anxiety, mood, sleep and quality of life at three months, with moderate-to-large effect sizes — a marked contrast to the RCTs' null results.
Several factors explain this gap. RCT samples are small and drawn from less severely unwell, non-clinical populations, meaning most participants would be ineligible for UK prescription. Strict exclusion criteria would rule out the majority of real-world patients, particularly those with comorbidities or prior cannabis use — the very people most likely to benefit. RCTs have also almost exclusively tested single-dose CBD isolates, unlike the varied, sustained regimens used in practice. Safety findings, however, align closely across both approaches.
We conclude that RCTs, as currently designed, are ill-suited to evaluating CBMPs for anxiety, and that future research should prioritise representative, real-world study designs over further trials of this kind.
To access the full publication from the Drug Science Policy and Law journal, please see below:
