The Evidence Gap: Why Outcomes Data Will Make or Break Cannabis Medicine

A Reckoning Built on Numbers

The medical cannabis industry across Europe is entering a new and more demanding phase. For several years, the dominant argument has been about access — getting patients to prescriptions, getting prescriptions written, and getting regulators to expand approved indications. That argument, in markets like the United Kingdom, has largely been won. The next one will be fought on clinical outcomes, and early signs suggest the sector is poorly prepared for it.

In March 2025, The Lancet Psychiatry published a systematic review of cannabinoids for mental health disorders. Its conclusion was blunt: the available randomised controlled trial evidence does not support routine prescribing for conditions including anxiety, PTSD, depression, or psychotic disorders. The Wilson review, a key study referenced in the Lancet paper, screened nearly 5,800 studies, included 54 randomised controlled trials, and found no significant benefit across the major mental health categories. Evidence certainty was rated very low or low under the widely used GRADE framework.

The response from the UK media was swift and critical. Calls for tighter prescription oversight followed, with headlines focusing on the scale of the market: private cannabis prescriptions in the UK more than doubled from 283,000 in 2023 to approximately 659,000 in 2024, according to figures from NHS Business Services Authority freedom of information data. The prescribing curve had simply outrun the evidence base.

Real-World Data Is Not Enough — But It Matters

The industry's response has leaned on real-world evidence, and that evidence is not without value. Drug Science, which operates the UK's largest non-profit medical cannabis registry with more than 4,500 patients followed for up to five years, has reported significant symptom reductions at three months in patients with comorbid depression and PTSD. A two-year case series of nearly 700 depression patients drawn from the UK Medical Cannabis Registry, published in the Journal of Affective Disorders, found significant reductions in depression severity at multiple timepoints. A separate prospective cohort published in late 2025 found that patients above clinical anxiety and depression thresholds dropped from over 80% and 76% respectively at baseline to below 50% within three months.

These are meaningful signals. They are not, however, the same as randomised controlled trial evidence, and the distinction is not merely academic. Registry and cohort data reflects patients who sought out treatment, paid privately, and remained in care — a self-selecting group that would not qualify for most trial designs. As Drug Science's Dr Anne Schlag has noted, many patients seeking cannabis prescriptions carry multiple concurrent diagnoses, precisely the kind of complexity that RCTs are designed to exclude.

The honest position is that real-world data and randomised trial evidence answer different questions. The former tells you what happens to patients who actually seek this treatment in clinical practice. The latter tells you whether the treatment itself is the cause of any improvement. Both are necessary. Right now, the field has an abundance of neither.

A Data Infrastructure Problem

Compounding the evidentiary challenge is a more basic problem: the quality of prescription data currently in circulation is poor. An analysis of NHS Business Services Authority records covering more than 132,000 prescription entries found that private cannabis prescriptions are manually recorded from handwritten prescriptions with no standardised naming convention. Product match rates fell as low as 77.5% in 2022, meaning nearly one in four prescription records from that year could not be reliably assigned to a specific medicine. Much of the mainstream media reporting on the UK cannabis market — including a high-profile investigation published in April 2025 — has drawn on datasets with these known limitations.

This is not a minor administrative inconvenience. It means that the figures used to drive policy debate, fuel media investigations, and inform regulatory decisions are built on a fractured foundation. Clinics and registries that have invested in rigorous, longitudinal, standardised data collection are not simply doing good science — they are building the only credible defence against regulatory pressure that is likely to intensify.

What This Means for Malta

Malta's cannabis framework is structured differently from the UK's private prescription market, but the underlying lesson is directly applicable. ARUC oversees a system in which Cannabis Harm Reduction Associations supply members with cannabis for personal use, rather than a clinical prescription model. However, as the Maltese framework matures and questions of medical use, product standards, and public health outcomes inevitably arise, the absence of systematic outcomes data will become a liability.

At least one CHRA operating in the central region of Malta has, according to members familiar with its practices, begun keeping informal records of member feedback and consumption patterns. This is a starting point, but it falls well short of the kind of structured, longitudinal data collection that would allow Malta to contribute meaningfully to the European evidence base or to defend its regulatory model against future scrutiny.

ARUC has a clear interest here. An authority tasked with responsible use cannot meaningfully assess responsibility without data. If Malta wants to position itself as a thoughtful, evidence-informed jurisdiction rather than simply a permissive one, investing in systematic outcome monitoring — even at a modest scale — is a practical next step. The UK's experience shows what happens when a market scales without building that infrastructure first.

The evidence gap in cannabis medicine is real and acknowledged even by advocates. The question for regulators, clinicians, and industry participants across Europe is not whether to fill it, but how quickly and rigorously they are willing to try. For Malta, the advantage is timing: the framework is young enough that good data habits can still be built in from the beginning, rather than retrofitted after the fact.

Featured image: Photo by MART PRODUCTION on Pexels