The UK’s medical cannabis market has long been a locked box when it comes to reliable data. The fact that it is almost entirely private means that the data is largely proprietary, forcing us to piece together and model the market from snippets given to us by individual clinics or to parse large, unreliable data sets attained through Freedom of Information requests.
Despite promises of a central ‘patient registry’ spanning over four years, no such official database yet exists. Several unofficial registries, such as the UK Medical Cannabis Registry, established in 2019 by Curaleaf Clinic physicians, have attempted to fill this gap, but ultimately leave analysts to make educated guesses.
The adverse impact on not only industry players, but on the wider healthcare system and both existing and potential patients should not be understated. It inhibits research into the wider impact of readily accessible medical cannabis, keeps the industry largely in the shadows, and means the patches of data that are made public are more prone to being misunderstood or misrepresented.
Slowly, this dynamic is beginning to change. With each fragment that is released, another brick in the wall is removed, and the picture becomes clearer.
In March, The Times published a story based on data obtained through an FOI request, shedding some light on prescription practices in the UK for the first time. We’ve now had the opportunity to examine that very data set, and below we’ll attempt to explain exactly what it can reliably show us, and perhaps more importantly, what it cannot.
What the data shows
Our sister publication, Business of Cannabis obtained the NHSBSA’s FOI-03587 dataset and subjected it to our own independent analysis. The following findings are drawn directly from that data. For each, we set out what the figures show, the caveats the NHSBSA attaches to their interpretation, and the additional context provided by the Medical Cannabis Clinicians Society.
Rapid growth
Between 2019 and 2024, the volume of cannabis-based medicinal products dispensed privately in England grew from 278 items to 663,130, a 2,382-fold increase over five years.
*2025 provisional Jan–May only. Source: NHSBSA FOI-03587
Monthly volumes mirror this trend. In early 2020, fewer than 400 items were dispensed every month. In the first five months of 2025, this monthly average had reached approximately 85,000. The 2024 total exceeded the combined total of every preceding year.
The NHSBSA states in its FOI response that the data is captured for monitoring purposes and is explicitly unvalidated. CBPM prescriptions are identified through a manual review of prescriptions filed as ‘unspecified drugs,’ not through an automated system, and completeness cannot be independently verified.
Furthermore, the 2025 figures are likely to change. Prescriptions are logged against the date written rather than the date submitted, and in our experience, these figures change substantially as reporting delays are realised.
Recognising the ‘significant growth in prescribing’, the MCCS says it sees this as a reflection of ‘sustained patient demand’, rather than a supply-driven phenomenon.
This, it says, is likely prevalent in ‘people with chronic, treatment-resistant conditions such as pain, anxiety and children with resistant epilepsy who have not found adequate relief from conventional therapies.’
It added: “Many patients are turning to private care because they are unable to access timely support through the NHS, with long waiting lists, including around 1.7 million people waiting for mental health services, contributing to this shift. Clinicians are responding where standard treatments have not worked, rather than as a first-line approach.”
Concentrated prescribing to a handful of doctors
Since 2019, nine prescriber identifiers have collectively accounted for more than half of all CBPM items recorded in England. The top ten combined account for 805,255 items, 52.2% of the all-time total of 1.54 million.
Thirty-seven identifiers account for 80% of all items ever dispensed. At the other end of the distribution, 121 of the 314 identifiers in the dataset have issued fewer than 100 items in total across the entire period.
In 2024, the concentration was more pronounced still. Three identifiers each dispensed 50,000 items or more, accounting for 31.6% of that year’s total between them.
Eleven further identifiers in the high-volume tier (10,000–50,000 items) contributed 37.5%. The 113 identifiers at the lowest tier, each dispensing fewer than 500 items across the full year, collectively accounted for under 2% of 2024 volume.
Furthermore, this concentrated prescribing dynamic, often attributed by the industry to the severe lack of specialists able to prescribe, has been declining since 2022.
The Gini coefficient, a measure of inequality across the prescriber base, peaked at 0.861 in 2022 and fell to 0.850 in 2024 and 0.775 in the first five months of 2025. The top single identifier’s share of annual volume fell from 14.9% in 2023 to 12.2% in 2024 and 10.8% in the provisional 2025 data. The number of active prescriber identifiers rose from 116 in 2022 to 200 in 2024, with 77 new entrants recorded that year alone.
Market concentration (Gini coefficient), 2019–2025
There are two important limits to what these figures can show. First, it remains unclear whether each prescriber identifier represents a single individual clinician or whether an identifier may in some cases correspond to a group practice or organisational account.
The NHSBSA told Business of Cannabis: “In most cases, a unique PIN is issued to an individual clinician and so a unique PIN should represent a single individual clinician.
“However, there are some group PINs in existence. We have not carried out any analysis on whether any of the PINs disclosed as part of FOI-03587 relate to a group.”
As such, where a group PIN appears in the data, items attributed to it cannot reliably be taken to represent the prescribing activity of a single individual.
Second, the data captures total item volume but cannot distinguish between initial assessments and ongoing repeat prescribing for established patients.
For example, a high-volume identifier may be issuing large numbers of routine repeat prescriptions within established care relationships, a pattern that looks identical in the data to high-volume new patient activity.
For this reason, the MCCS cautions against ‘drawing conclusions about the quality or safety of care from prescribing volume data alone,’ adding that ‘aggregate or anonymised figures do not capture the clinical context, models of care, or governance processes in place.’
The society is unambiguous that ‘item-level prescribing volume is not a reliable indicator of clinical activity, workload, or whether appropriate standards are being met.’
The scale of leading prescribers
The highest-volume identifier in the dataset has issued 172,755 items since first appearing in the data in early 2020, 11.2% of the all-time total.
In the first five months of 2025, that identifier recorded 45,947 items, averaging approximately 9,189 items per month. The peak single month in the dataset for this prescriber was April 2025, when 9,697 items were recorded.
Once again, it’s worth noting that the 2025 figures are likely to change due to reporting delays, with the most recent months likely to see the largest changes.
A critical caveat here is that an item is the number of times a product appears on a prescription form, not the number of prescriptions issued and not the number of patient consultations.
“The term ‘items’ refers to the number of times a product appears on a prescription form, not the quantity prescribed. One item should not be read as one prescription issued. We do not hold any data on patient consultations,” the NHSBSA told us.
A single appointment generating a prescription for three cannabis products would produce three items in this dataset. The raw item totals in FOI-03587 cannot be used to estimate how many patients received treatment, how many consultations took place, or how frequently any individual was seen.
The MCCS added: “Prescribing data reported as ‘items’ does not equate to the number of patients treated or consultations delivered. On average, this is around two items per patient, per prescription, though this varies.
“Some clinicians provide one prescription containing multiple items; others issue separate prescriptions for each item. These differences can significantly affect how prescribing volume is recorded and interpreted.”
At two items per patient per prescription, the MCCS’s stated average, the leading identifier’s 45,947 items in the first five months of 2025 would imply approximately 22,974 patient-prescriptions over that period, or roughly 4,595 per month.
The NHSBSA has confirmed that the supplementary calculations circulated alongside the dataset — including the methodology note describing an assumed working week of 20 days and eight hours — were not part of its FOI response. The authority that collects and holds the data did not produce them and is unable to verify them. Any arithmetic extrapolating a per-minute or per-hour prescribing rate from the item totals is therefore a third-party calculation, not one the NHSBSA has validated or endorsed. The pace implied by those figures should be read in that light.
On whether that volume is consistent with appropriate assessment, peer review, and monitoring under MCCS guidance, it said noting that the ‘responsibility for ensuring safe and appropriate prescribing sits with the individual prescriber and the clinic.’
Comparing medical cannabis and street cannabis
The Times reported that the strongest medical cannabis available in the UK, a product named ‘Space Cake’, contains 34% THC, more than double the 14–16% typically found in street cannabis. The comparison is presented as an indicator of risk.
Asked whether the NHSBSA holds data at this level of granularity for CBPM prescribing, it told Business of Cannabis: “This data has been previously published in response to FOI-03623. However, this data cannot be validated. This is because, as part of our manual processing exercise, we capture product descriptions as they are written on the prescription form.
“Cannabis-based products for medicinal use are unlicensed and are not included in the NHS Dictionary of Medicines and Devices. This means there is no standardised naming convention for the product names, strength or volume. This may lead to duplication of some product names, strength or volume depending on variations within how the product was prescribed and/or recorded.”
On the comparison itself, the MCCS notes that medical cannabis and illicit cannabis operate in entirely different clinical contexts. Prescribed doses are titrated from low starting points under specialist supervision, products are quality-controlled, and cannabidiol can mitigate some of the psychoactive effects of THC.
The MCCS Good Practice Guidance, updated in May 2026, addresses this directly, recommending that prescribers avoid ‘terminology derived from recreational cannabis markets’ on the grounds that such names ‘have very limited clinical meaning, do not support professional communication and may undermine acceptance of CBMPs as legitimate medicines.’ The guidance notes that ‘it is good to see a reduction in this practice over the last 12 months across many parts of the sector.’
Beyond the data
Within 48 hours of The Times article publishing, two further FOI requests were submitted to the NHSBSA. The first, FOI-03715, asked the authority to confirm whether a named doctor was among the ten high-volume prescribers and, if so, to name the other nine, or failing that, to name the clinics involved.
The second, FOI-03716, asked for the names, GMC registration numbers, and prescriber codes behind the 10 pseudonymised identifiers in the dataset.
Both were refused by the NHSBSA on the grounds of personal data protection under Section 40(2) and prejudice to the effective conduct of public affairs under Section 36.
“The data that NHSBSA holds shows volumes of prescriptions issued by named prescribers. On its own, the data we hold cannot provide any wider context about the prescribing and/or its appropriateness.”
The authority that collected the data and owns its caveats concluded that it could not fairly be used to identify individuals publicly. That position is consistent with every limitation we have set out in the data section above,
The Advisory Council on the Misuse of Drugs has been commissioned to review the use and availability of cannabis-based products for medicinal use, a process referenced by the NHSBSA in its FOI refusal letters and one that will, in time, produce recommendations that carry regulatory weight.
The MCCS’s latest Good Practice Guidance acknowledges that ‘the absence of mandatory training and consistently applied clinical standards has led to variation in practice across the sector.’
Whether that translates into enforceable standards, and on what timeline, is the question the sector is waiting for the ACMD review to begin to answer.
This piece was originally published by Business of Cannabis and is reprinted here with permission.
