Chronic pain is one of the leading causes of disability worldwide. Many people with chronic pain do not get enough relief from current treatments, and some treatments carry risks. A growing number of people in the UK are prescribed cannabis-based medicinal products (medical cannabis) for chronic pain. However, the evidence for these medicines is still limited.
A study published in The Journal of Clinical Pharmacology in 2026 looked at 3,744 people in the UK who were prescribed medical cannabis for chronic pain. The researchers followed their pain scores for up to 2 years. They wanted to find out whether people’s pain changed in different ways over time, and which personal factors were linked to each pattern.
Key Takeaways:
- The study. Researchers analysed data from 3,744 people in the UK Medical Cannabis Registry whose main reason for treatment was chronic pain. Each person had been prescribed medical cannabis by a specialist doctor. Their pain was tracked over 24 months.
- Five patterns. Using a statistical method called clustering, the team found five distinct patterns of change in pain severity over time.
- Not everyone had a meaningful improvement. In the largest group (956 people, 25.53%), pain severity at 24 months was almost the same as at the start. Three groups, together 1,842 people (49.20%), had improvements that lasted for 24 months.
- Personal factors mattered more than products. Sleep quality, anxiety and type of pain at the start were more closely linked to a person’s pattern than the product or dose prescribed.
- Side effects. 538 people (14.37%) reported at least one side effect. Most reports were mild or moderate. Fatigue, dry mouth, lethargy and sleepiness were the most common.
- What it cannot show. There was no comparison group. The study cannot show that medical cannabis caused the changes seen.
What is Chronic Pain?
Chronic pain is pain that lasts or keeps coming back for more than 3 months. It affects more than 30% of people worldwide. In the UK, just under 28 million adults are estimated to live with chronic pain. Between 10.4% and 14.3% of UK adults have pain that is moderately to severely disabling.
Chronic pain often affects more than the body. It is closely linked with anxiety, low mood and poor sleep. Each of these can make the others worse and lower a person’s quality of life.
Recovery from chronic pain is uncommon. In a large Norwegian study called HUNT, people with moderate to severe chronic pain had only an 8% chance of recovering over 1 year. Among those who did recover, only 52% were still pain-free at the next check. The longer the pain had lasted, the less likely recovery became.
How is chronic pain usually treated in the UK?
Treatment aims to reduce pain and its impact on daily life. It also aims to help with related problems such as anxiety and poor sleep.
For chronic primary pain, which is chronic pain with no clear underlying cause, the National Institute for Health and Care Excellence (NICE) recommends (NG193):
- Exercise programmes. Supervised group exercise to improve movement and reduce pain.
- Psychological therapies. Such as cognitive behavioural therapy (CBT) or acceptance and commitment therapy (ACT), to help people manage the physical and emotional impact of pain.
- Acupuncture. As a single course, in some cases.
- Some antidepressants. These may be considered, even when a person does not have depression.
NICE advises that several common medicines should not be started for chronic primary pain. These include opioids, non-steroidal anti-inflammatory drugs (NSAIDs), paracetamol and gabapentinoids (such as pregabalin and gabapentin). This is because there is little evidence that they help, and they can cause harm or dependence.
Chronic pain caused by another condition, such as arthritis or nerve damage, is managed according to the condition. For example, NICE has separate guidance for nerve pain (neuropathic pain).
Can medical cannabis be prescribed for chronic pain in the UK?
Yes, in some cases. Since November 2018, specialist doctors on the General Medical Council (GMC) Specialist Register have been able to prescribe medical cannabis. Chronic pain is the most common reason people are prescribed it.
The medical cannabis products prescribed for chronic pain are unlicensed. This means they have not been through the full approval process for this use. NICE does not recommend medical cannabis for chronic pain (NG144). As a result, they are not routinely available on the NHS for chronic pain. Most people who are prescribed it receive treatment privately.
Medical cannabis contains compounds from the cannabis plant called cannabinoids. The two main ones are delta-9-tetrahydrocannabinol (THC) and cannabidiol (CBD).
What does the wider research say?
Clinical trial evidence for medical cannabis in chronic pain is limited and mixed. Different reviews have reached different conclusions:
- A review of 32 trials found that non-inhaled medical cannabis led to a small improvement in pain relief compared with placebo (Wang et al., 2021). A placebo is a dummy treatment with no active ingredient.
- A review of 36 trials concluded that the evidence neither supported nor ruled out a benefit of medical cannabis for pain (Fisher et al., 2021).
- A review of 65 trials found that cannabinoids reduced chronic pain, but by less than the smallest change patients would notice. The authors judged that the harms outweighed the benefits (Barakji et al., 2023).
Across these reviews, any average benefit was small. However, averages can hide big differences between people. Some people may improve a lot while others do not improve at all. This study set out to look at those differences.
What is the UK Medical Cannabis Registry?
The UK Medical Cannabis Registry was set up in December 2019 by Curaleaf Clinic (then called Sapphire Medical Clinics). It collects information on the health and safety of people prescribed medical cannabis in the UK and Crown Dependencies. Patients give written consent to take part. The Registry has approval from a research ethics committee.
People in the Registry complete questionnaires about their symptoms online. They do this before starting treatment and then at regular points afterwards.
An earlier study from the Registry looked at quality of life across all conditions. It found that people followed different patterns over time, including some groups who did not improve. This new study looked at whether the same was true for pain.
What did this study do?
This was an observational study. The researchers analysed information that had already been collected. They did not give treatment as part of a trial, and there was no comparison group.
Who was included?
The study included adults in the Registry who:
- had chronic pain as their main reason for treatment
- had been prescribed unlicensed medical cannabis by a specialist doctor
- had completed questionnaires before starting treatment
- had been in the Registry for at least 2 years by January 2025
At the first appointment, doctors recorded each person’s age, sex, weight, job, medical history, other conditions and current medicines. They also recorded smoking, alcohol use and whether the person had used cannabis before.
What did the researchers measure?
People completed questionnaires at the start and at 1, 3, 6, 12, 18 and 24 months. These measured:
- Pain severity and interference. The Brief Pain Inventory (BPI) scores how bad pain is, and how much it interferes with daily life, from 0 to 10.
- Types of pain. The Short-Form McGill Pain Questionnaire-2 describes different kinds of pain, including nerve pain.
- Overall pain. A pain scale from 0 (no pain) to 10 (worst pain imaginable).
- Quality of life. The EQ-5D-5L questionnaire.
- Anxiety. The Generalised Anxiety Disorder-7 (GAD-7) questionnaire.
- Sleep. The Single-Item Sleep Quality Scale.
The team also recorded side effects and changes in other prescribed pain medicines.
For each measure, the researchers set a level of change that patients would be likely to notice. This is called a minimal clinically important difference. For pain severity, it was a fall of at least 1 point on the 0 to 10 scale. They also counted larger improvements of at least 30% and at least 50%.
What is clustering?
Clustering is a statistical method that groups people whose results follow a similar pattern. Here, the computer grouped people by how their pain severity scores changed over the 24 months. It did this without being told in advance what the groups should look like. The researchers then looked at what the people in each group had in common at the start.
Who took part?
Of 34,563 people in the Registry, 3,744 met the criteria for this study.
- The average age was 46.
- 2,172 (58.01%) were men and 1,572 (41.99%) were women.
- 2,152 (57.48%) were using cannabis before starting treatment and 590 (15.76%) had used it in the past. 1,002 (26.76%) had never used cannabis.
The most common main diagnoses were chronic non-cancer pain (2,276 people, 60.79%) and nerve pain (600 people, 16.03%). Others included hypermobility, inflammatory arthritis, osteoarthritis, cancer pain, endometriosis and complex regional pain syndrome.
Fewer people completed questionnaires over time. At the start, 3,658 people (97.70%) gave a pain score. By 24 months, this had fallen to 1,309 (34.96%). The researchers used a statistical method called multiple imputation to estimate the missing scores. This is explained under limitations below.
What did the study find?
Did symptoms change across the whole group?
Yes. On average, every measure was better at each check-up than at the start. The table shows average scores at the start and at 24 months.
| Measure | At the start | At 24 months |
| Pain severity (0 to 10, lower is better) | 5.97 | 4.86 |
| Pain interference with daily life (0 to 10, lower is better) | 6.87 | 5.38 |
| Overall pain scale (0 to 10, lower is better) | 7.12 | 5.90 |
| Anxiety, GAD-7 (0 to 21, lower is better) | 8.51 | 5.29 |
| Sleep quality (0 to 10, higher is better) | 3.80 | 5.73 |
| Quality of life, EQ-5D-5L index (higher is better) | 0.30 | 0.53 |
These figures include estimated values for people who did not complete questionnaires. Among the 1,309 people who did complete the pain questionnaire at 24 months, pain severity fell from 5.86 to 4.55 on average.
How many people had a meaningful improvement?
At 24 months, just over half of people had an improvement in pain severity that patients would be likely to notice. Fewer people had larger improvements.
| Improvement at 24 months | Pain severity | Pain interference |
| At least 1 point (noticeable) | 2,045 of 3,744 (54.62%) | 2,350 of 3,744 (62.77%) |
| At least 30% (moderate) | 1,285 of 3,744 (34.32%) | 1,575 of 3,744 (42.07%) |
| At least 50% (substantial) | 557 of 3,744 (14.88%) | 794 of 3,744 (21.21%) |
Other measures at 24 months:
- Quality of life. 2,872 people (76.71%) had any improvement in their quality of life score.
- Anxiety. 1,587 people (42.39%) had a noticeable improvement.
- Sleep. 1,464 people (39.10%) had a noticeable improvement.
What were the five patterns of pain?
Clustering found five groups, labelled A to E. Each group ended with a lower average score than it started with. But the size of the change, and whether it lasted, varied a lot.
| Group | Number of people | Average pain severity at the start (0 to 10) | At 24 months | Pattern |
| A | 956 (25.53%) | 5.85 | 5.79 | Small early improvement, then back close to the starting level |
| B | 946 (25.27%) | 5.10 | 4.44 | Clear improvement by 3 months that partly faded |
| C | 679 (18.14%) | 7.72 | 6.30 | Highest starting pain. Slow, modest improvement |
| D | 675 (18.03%) | 6.86 | 4.03 | Largest improvement, which lasted |
| E | 488 (13.03%) | 4.26 | 2.57 | Mildest starting pain. Quick improvement in the first month, which lasted |
In groups C, D and E, the average improvement at 24 months was more than 1 point. Groups A and B, together just over half of all people, had an early improvement that did not last. By 24 months, their average change was less than 1 point.
The largest group, group A, showed almost no change by 24 months. This shows that an early improvement after starting treatment does not always continue.
What was linked to each pattern?
The researchers compared what people in each group had in common at the start. They found that:
- Poor sleep was linked to being in groups C and D, which started with higher pain. It was linked to being less likely to be in group E.
- Severe anxiety and being prescribed opioids at the start were linked to being in group C.
- Being unemployed was linked to being in groups C and D.
- Cancer pain or nerve pain as the main diagnosis was linked to being in group E.
- Women and older people were less likely to be in group E.
- Past cannabis use was linked to being less likely to be in group D.
The product type and dose prescribed were not strongly linked to which group a person was in. The authors say a person’s starting characteristics were more closely linked to their pattern than the products prescribed.
These factors only partly explained which group people ended up in. They cannot be used to predict how any one person will respond.
The authors urge caution about the cancer pain finding. People with cancer pain were the least likely to complete questionnaires, so their results relied more on estimated values. Other cancer treatments may also have played a part.
Were any factors linked to a noticeable improvement?
Yes. Across the whole group, poor sleep and higher anxiety at the start were the factors most often linked to a noticeable improvement at 24 months:
- People with poor sleep at the start had about twice the odds of a noticeable improvement in pain severity and overall pain.
- People with severe anxiety at the start had higher odds of a noticeable improvement in pain severity.
- People prescribed opioids or gabapentinoids at the start had higher odds of improvement in quality of life.
- People whose main diagnosis was cancer pain had lower odds of improvement in pain interference and overall pain.
The authors suggest two possible reasons for the sleep and anxiety findings. People with the worst sleep and anxiety have the most room to improve. And better sleep and lower anxiety may in turn reduce how much pain interferes with daily life.
Did the use of other pain medicines change?
The average prescribed daily dose of opioid painkillers was lower than at the start from 3 months onwards. Among people prescribed opioids at the start, 302 (22.66%) had a reduction of at least 28.2% by 24 months. This is a level considered clinically important. Prescribed doses of gabapentinoids were lower from 6 months onwards.
These findings should be treated with caution. Medicine changes were only counted if they were recorded. The authors note that similar findings in other observational studies are based on very low-certainty evidence. They have not yet been confirmed in randomised controlled trials.
Important: Do not stop or change any prescribed pain medicine without talking to your doctor. Stopping some medicines suddenly, including opioids and gabapentinoids, can cause withdrawal symptoms.
What side effects were reported?
538 people (14.37%) reported at least one side effect. In total, 4,898 side effects were reported. Of these:
- 2,027 (41.38%) were mild
- 2,143 (43.75%) were moderate
- 720 (14.70%) were severe
- 8 (0.16%) were life-threatening or disabling
The most frequently reported side effects were fatigue (400 reports), dry mouth (303), lethargy (300) and sleepiness (300). Headache, insomnia, nausea, dizziness, difficulty concentrating and constipation were also among the 10 most common.
Some groups were more likely to report side effects. These included people aged over 50, women, people with hypermobility as their main diagnosis, and people on the highest CBD doses. People already using cannabis before treatment and people with cancer pain were less likely to report side effects.
Why might these results differ from clinical trials?
The improvements in this study were larger and lasted longer than those reported in reviews of clinical trials. The authors give several possible reasons:
- Expectations. Everyone knew they were receiving medical cannabis. Expecting to feel better can affect how people rate their symptoms.
- Who chose treatment. Most people were already using cannabis before treatment. People who had found it helpful in the past may be more likely to seek a prescription.
- Flexible treatment. Doctors could adjust the product and dose for each person. Most trials use fixed doses.
- Regression to the mean. People often seek new treatment when their symptoms are at their worst. Symptoms can then improve naturally, regardless of the treatment.
The authors also looked at why early improvements faded in groups A and B. Over the study, the THC doses prescribed tended to rise. This could reflect the body getting used to THC over time, which is called tolerance. It could also reflect doses being increased because the benefit was not enough, or the natural ups and downs of chronic pain.
What are the limitations of this study?
The authors point out several important limitations:
No comparison group. No one received a placebo or standard care instead. Without a comparison, it is not possible to say how much of the change was due to medical cannabis.
- Observational design. The study cannot show cause and effect.
- Missing data. Only about 1 in 3 people gave a pain score at 24 months. The researchers estimated missing scores using multiple imputation. However, people who stopped treatment because it did not help, or because of side effects, may be less likely to have completed questionnaires. This could make results look better than they are.
- Possible under-reporting. If no side effect or medicine change was recorded, the study assumed none had happened. This may underestimate both.
- One clinical service. Everyone was treated through one clinic. The results may not apply to other settings.
- Broad diagnosis groups. Many different pain conditions were grouped together. This may hide differences between them.
- Pain severity only. The groups were based on pain severity. They do not capture other effects, such as on daily life or quality of life.
The authors note that the lasting improvements in groups C, D and E are larger than would usually be expected from the natural course of chronic pain. However, because there was no comparison group, they cannot say that medical cannabis caused these improvements.
What does this study mean, and what does it not mean?
In this large UK group, people’s pain changed in very different ways after starting medical cannabis. About half had improvements that lasted for 2 years. About half had early improvements that faded or were very small. Most side effects were mild or moderate, but some were severe.
Sleep, anxiety and type of pain at the start were more closely linked to these patterns than the specific product or dose. The authors say that an early improvement cannot be assumed to last. They suggest that response to treatment should be reviewed regularly over time.
The study does not show that medical cannabis work for chronic pain. It does not show that they will help any one person. Randomised controlled trials, or new ways of using real-world data, are needed to find out the true effect of medical cannabis.
Frequently asked questions
Can medical cannabis be prescribed for chronic pain in the UK?
Yes, in some cases. A specialist doctor on the GMC Specialist Register can prescribe medical cannabis, usually when other treatments have not helped enough. The medical cannabis products used for chronic pain are unlicensed. NICE does not recommend them for chronic pain, so they are not routinely available on the NHS for this purpose.
Does this study show that medical cannabis works for chronic pain?
No. This was an observational study with no comparison group. On average, pain, anxiety, sleep and quality of life scores were better at 24 months than at the start. However, the study cannot show that medical cannabis caused these changes. Expectations, natural changes in pain and missing data may all have played a part.
What were the five pain patterns?
The study grouped 3,744 people by how their pain severity changed over 24 months. In the largest group (956 people), pain was almost unchanged by 24 months. In a second group (946 people), early improvement partly faded. In three groups (1,842 people), improvements lasted to 24 months. One of these started with the highest pain and improved only modestly.
What side effects were reported?
538 of 3,744 people (14.37%) reported at least one side effect. Of 4,898 reported side effects, most were mild (41.38%) or moderate (43.75%). 720 (14.70%) were severe and 8 (0.16%) were life-threatening or disabling. The most frequently reported were fatigue, dry mouth, lethargy and sleepiness.
What does NICE say about medical cannabis for chronic pain?
NICE advises doctors not to offer THC, CBD with THC, nabilone or dronabinol for chronic pain in adults. It advises that CBD should only be offered for chronic pain as part of a clinical trial (NG144). People who started medical cannabis for chronic pain on the NHS before this guidance can continue until they and their doctor agree to stop.
Who carried out the study?
The study was carried out by researchers at Curaleaf Clinic and Imperial College London. Curaleaf Clinic runs the UK Medical Cannabis Registry. The lead author, Dr Simon Erridge, is Research Director at Curaleaf Clinic. All but one of the other authors work for Curaleaf Clinic or Curaleaf International. The remaining author, Dr Pedro Mediano, is based at Imperial College London.
Considering your treatment options?
Living with chronic pain can affect every part of life. If your pain is not well controlled, speak to your GP or pain team first. They can review your current treatment and explain which options may suit you. Exercise programmes and psychological therapies are often recommended, sometimes in combination. The charities Pain UK and the British Pain Society also offer advice and support.
Medical cannabis products for chronic pain are unlicensed in the UK. A specialist doctor may consider them only after a full assessment of your medical history and previous treatments. To find out whether a specialist consultation may be appropriate for you, you can complete Curaleaf Clinic’s online eligibility check. Any decision about treatment is made by a specialist doctor on an individual basis.
