Medicinal cannabis for chronic pain: what the evidence really says

Pain is one of the most common reasons people seek medicinal cannabis — in NZ's 2025 drug-trends data, pain was a reason for use in around 63% of medicinal users. Given that, it's worth looking honestly at what the evidence actually shows. The short version: there is a large body of research, but the demonstrated benefit is modest, and cannabis sits as a last-line add-on rather than a first choice.

Chronic non-cancer pain

This is the broad category most patients fall into. The evidence here is best described as mixed, limited-to-moderate: a large evidence base, but a small average effect.

• A major 2018 review (Stockings, in the journal PAIN) found that 30% pain reduction was achieved in about 29% of patients on cannabinoids versus about 26% on placebo — a small difference, giving a "number needed to treat" of around 24. The "number needed to harm" for adverse events was much lower, at around 6.
• A 2021 BMJ analysis estimated you'd need to treat about 10 people to achieve a modest (1 cm on a 10 cm scale) reduction in pain.

In plain terms: cannabis helps some people with chronic pain, but on average the effect is small, and side effects are common. That's why it's positioned as a last-line add-on, and why the BMJ guidance leaned toward non-inhaled routes.

Neuropathic (nerve) pain

For nerve pain specifically, the evidence is limited and low-quality, and the harms loom larger.

• A 2018 Cochrane review (Mücke) found 30% pain relief in about 39% on cannabinoids versus 33% on placebo — a "number needed to benefit" of around 11. But nervous-system adverse events occurred in about 61% versus 29%, a "number needed to harm" of just 3.
• The reviewers concluded the potential benefits "might be outweighed by harms."
• The International Association for the Study of Pain does not endorse routine use, and the UK's NICE recommends against it outside trials.

So for neuropathic pain, cannabis is generally reserved behind established options like gabapentinoids, duloxetine and amitriptyline.

Specific pain conditions

• Fibromyalgia: the evidence is limited and insufficient. A 2016 Cochrane review found only two small trials and no convincing unbiased evidence, with poor tolerability. It's an adjunct after standard approaches (exercise, duloxetine or pregabalin, CBT).
• Endometriosis and chronic pelvic pain: limited and emerging — observational surveys report self-reported pain reduction, but there are no randomised trials. (And cannabis should be avoided in pregnancy or when trying to conceive.)
• Migraine and headache: limited and emerging, mostly from retrospective data; controlled trials are only now beginning. A real caution here is medication-overuse (rebound) headache with frequent acute use.

What this means for patients

A few honest takeaways:

• Cannabis is not a first-line pain treatment in New Zealand, and it isn't a guaranteed solution. For many, the benefit is modest or absent.
• It's an add-on, generally tried after standard treatments, and it works best as part of a broader pain-management plan that may include exercise, physiotherapy, psychological approaches and other medicines.
• Side effects are common, so the "start low, go slow" approach and realistic goals matter.
• A time-limited trial with clear measures is the sensible way to find out if it helps you — keep a pain-and-side-effect diary, and be willing to stop if there's no clear benefit.
• Product choice matters. Balanced THC:CBD or THC-containing products underpin most of the pain evidence; non-inhaled routes were favoured in the BMJ guidance.

Why the gap between research and reports?

Patients often ask why so many people report strong relief when the trial data shows only a small average effect. There are several honest explanations. Averages hide variation — a modest average benefit can include some people who respond well and many who don't, and those who respond are naturally the ones who keep using it and talk about it. Expectation and the placebo response are powerful in pain, which is subjective. Chronic pain also fluctuates, so improvements that would have happened anyway can be attributed to a new treatment. And online testimonials are self-selected. None of this means cannabis doesn't help anyone — clearly it helps some — but it explains why enthusiastic anecdotes and cautious trial results can both be true at once, and why a personal trial with objective tracking is the only real way to know if it helps you.

How to trial it sensibly

If you and your prescriber decide to try cannabis for pain, a structured approach gives the clearest answer: set a specific goal (for example, a measurable reduction in pain or improvement in function), keep a daily diary of pain scores, doses and side effects, titrate slowly, and review at a set point — say six to eight weeks. If there's a clear, worthwhile benefit that outweighs side effects, continue at the lowest effective dose; if not, stop. This disciplined approach respects both the possibility of benefit and the reality that, for many, the effect is small.

The bottom line

For chronic pain, medicinal cannabis is a legitimate but modest option — helpful for some, ineffective for others, and rarely dramatic. The research supports trying it as a last-line add-on, with realistic expectations and careful attention to side effects, rather than as a first or sole treatment. Approached that way, it can be a worthwhile part of a comprehensive pain plan; sold as a cure, it will disappoint.

Last reviewed 12 October 2026 — education, not medical advice. Sources: Stockings 2018 (PAIN); Cochrane neuropathic pain (Mücke 2018); IASP position statement.