“My back hurts, I should get it scanned.”
It is the most natural thought in the world.
Unfortunately, for most back and neck pain, the scan won’t tell us much. It rarely changes what we do.
For most back and neck pain, though, the scan won’t tell us much. It rarely changes what we do. And in several studies, people who were scanned early did worse rather than better.
I know that sounds counterintuitive, so let me explain.
What a scan is good at
Think about your computer packing up. Your first move isn’t to unscrew the case and stare at the motherboard. Almost always the fault is in the software, not the hardware you can see.
A scan of your spine is a photograph of the hardware.
What it usually can’t show is why a nerve has become irritable. A scan can sometimes catch a nerve root that is squashed or inflamed, but most of the time the nerve that hurts looks ordinary in the picture. It can’t show a pathway in the spine that is firing more than it should. It can’t show how your tissues are behaving under load, or how your system is regulating pain today versus last Tuesday.
Those things are real and physical, and as far as we can tell they are what usually drives persistent back pain. We just don’t have a camera that photographs them yet.
That is why NICE, in its guideline on low back pain and sciatica, tells non-specialist services not to routinely offer imaging, and to consider it in specialist settings only if the result is likely to change management. The reasoning in the accompanying quality standard is blunt: the findings are usually common, not necessarily related to symptoms, and can cause unnecessary distress and further referrals for things that turn out not to matter. There is no equivalent NICE guideline for neck pain, so for the neck I am leaning on the data below and on general principle rather than a UK document.
What a scan finds in pain-free people
Here is what a scan does show. Brinjikji et al. (2015) pooled 33 studies, 3,110 people in total, all of whom had scans and no back pain at all.
| On a scan of a pain-free person | Age 30 | Age 50 | Age 70 |
|---|---|---|---|
| Disc degeneration | 52% | 80% | 93% |
| Disc bulge | 40% | 60% | 77% |
| Worn facet joints | 9% | 32% | 69% |
These are modelled estimates rather than raw counts, so they are not “X out of 100 people scanned”. If you are 50, there is a decent chance you have a bulging disc right now.
Degeneration on a report is closer to grey hair or wrinkles than to damage.
The neck is the same story. Nakashima et al. (2015) scanned 1,211 people with no symptoms and found disc bulging in 87.6% of them, including most of the people in their twenties. The findings that genuinely carry weight, compression of the spinal cord and changes in its signal, were rare: 5.3% and 2.3%.
So finding one of these things on your scan tells us very little. It was probably there before your back started hurting, and it will still be there long after your back has settled down.
Why this can actually cause harm
Once you have seen a word like degeneration on a piece of paper with your name on it, it is very hard to unsee.
Farmer et al. (2021) surveyed 677 people across five countries on the ordinary terms in back scan reports. The best-understood term was “disc degeneration”, and only 35% said they understood it. Across all 14 terms, between 59% and 71% of people took the term to mean something serious was wrong, and 42% to 57% said it would make them frightened of moving.
But the outcome research points the other way. Chou et al. (2009) pooled six randomised trials, 1,804 people, comparing immediate imaging with usual care for back pain that had no signs of anything sinister. No difference in pain or function at three months, none at six to twelve. In one of those trials, Jarvik et al. (2003), the people who got a rapid MRI did no better than the people who got an X-ray, but ten of them had spinal surgery against four in the X-ray group.
Early imaging is associated with worse outcomes in several studies, and with more injections, surgery and medication afterwards. Those studies show an association, not a proven cause. The likeliest explanation is what happens next: the fixation, the fear of movement, the sense of a broken part that needs fixing.
When you should not wait
Call 999 now, and do not drive yourself anywhere, if you have:
- sudden weakness in an arm or leg, or sudden numbness down one side of your body, even if it passes
Go to A&E now, or call 999, if you have either of these:
- numbness in the saddle area, meaning around the groin, buttocks or inner thighs
- any new change in your bladder or bowel control
The window for treating what those can mean is measured in hours.
Contact your GP urgently, today, if you have either of these:
- weakness in a leg or arm that is getting worse over days
- numbness that is spreading
When we do refer for a scan
This is not a policy of never. Beyond the urgent list above, there are ordinary situations where we will refer on.
- Pain that hasn’t shifted despite doing the right things. Six to twelve weeks of consistent, sensible treatment is a reasonable point to reassess properly. Time on its own is not the reason to scan; the reason is that, at that reassessment, a scan looks likely to change what we do next. That is the NICE test: image when the result will change management.
- A history that raises a different question. A past cancer, unexplained weight loss or fever, feeling generally unwell with it, a recent fall or accident, night pain that rest does not ease, or long-term steroid use. Any of those alongside back pain means a prompt GP assessment, not a wait-and-see with us.
- Neck symptoms that behave like the spinal cord rather than a nerve. Clumsy hands, a change in how you walk, symptoms in both arms or both legs. If your walking or balance changes suddenly, or is getting worse quickly, call 999 or go to A&E. If it has crept in gradually, see your GP within days, or call 111 if you cannot get an appointment. Hilton et al. (2019) traced the route to diagnosis for degenerative cervical myelopathy in the UK and found it took, on average, 12.5 months from first symptoms to a cervical MRI, with some people deteriorating in the meantime.
One from the clinic
I saw a guy recently who came in with a scan already done, showing one vertebra sitting slightly forward of the one below it. Understandably, he arrived certain this was the cause.
Except his pain was in a slightly different place to where that would put it, and it behaved in a way that didn’t fit. So we largely set it aside and treated what was actually in front of us, which was a back that had become intolerant of bending and needed loading properly.
He got better. The vertebra is still exactly where it was.
The honest counter-case
One large UK trial, Gilbert et al. (2004), across 782 people in 15 hospitals, found that early imaging produced a small but real improvement in back pain scores at eight and 24 months. Treatment barely changed between the groups, so the gain seems to have come from reassurance, and those were people already referred to hospital. But it is there: for some people, in some settings, seeing the picture helps them settle.
There is also a live argument among pain researchers that “scans correlate poorly with pain” has been stretched into “scans have no diagnostic role”, leaving people with severe pain and a clean scan told there is nothing biologically wrong. I think that overstates what NICE says. But the harm it describes is genuine, and it is the thing I most want you to take from this.
This will change
Imaging is improving. There is research now looking at inflammation in nerves and at things conventional scans were never designed to pick up. If those tools mature into something that changes how we treat you, we will use them.
But that is the future, and this is now. Right now, a scan of the hardware is a blunt instrument for a software problem.
The thing I most want you to take away
A normal scan does not mean nothing is wrong with you.
If you have been told your scan is clear and walked away feeling dismissed, or quietly wondering whether people think you are making it up: your pain is real, and it has a real biological cause. The camera was just pointed at the wrong thing.
Frequently asked questions
Should I get an MRI for my back pain?
For most back pain without the warning signs listed above, no. The UK guideline is not to routinely image in non-specialist settings, and the trials that compared early scans with no scans found no difference in pain or function. If a proper reassessment finds that a scan would change the plan, that is when we refer.
My scan says I have a bulging disc. Is that why it hurts?
Possibly, but not necessarily. Disc bulges show up in around 60% of pain-free 50-year-olds and 77% of pain-free 70-year-olds. The report tells us it is there. How your pain behaves under load usually tells us more.
Why would you need an MRI for back pain?
Mostly when the result would change what we do next. That means pain that hasn’t shifted after six to twelve weeks of sensible treatment, or a history that raises a different question, like a past cancer or unexplained weight loss. The section above covers each case.
General guidance, not a diagnosis. Sudden weakness in an arm or leg, or sudden one-sided numbness: call 999. Numbness in the saddle area or any new change in bladder or bowel control: go to A&E now or call 999. Weakness getting worse over days, or spreading numbness: contact your GP urgently, today.
If you’re worried about your back or neck and wondering whether you should get it scanned, book an assessment and we can talk it through properly. Happy to look at your particular situation rather than the general case, sciatica included.
References
- NICE. Low back pain and sciatica in over 16s: assessment and management (NG59), recommendations 1.1.4 to 1.1.6. https://www.nice.org.uk/guidance/ng59
- NICE. Quality standard QS155, statement 2: referrals for imaging. https://www.nice.org.uk/guidance/qs155/chapter/quality-statement-2-referrals-for-imaging
- Brinjikji W et al. Systematic literature review of imaging features of spinal degeneration in asymptomatic populations. AJNR 2015;36(4):811–816. https://www.ajnr.org/content/36/4/811
- Nakashima H et al. Abnormal findings on MRI of the cervical spines in 1211 asymptomatic subjects. Spine 2015;40(6). https://doi.org/10.1097/BRS.0000000000000775
- Kasch R et al. Association of lumbar MRI findings with current and future back pain in a population-based cohort study. Spine 2021;46(19). https://doi.org/10.1097/BRS.0000000000004198
- Farmer C et al. Consumer understanding of terms used in imaging reports requested for low back pain. BMJ Open 2021;11(9):e049938. https://doi.org/10.1136/bmjopen-2021-049938
- Chou R, Fu R, Carrino JA, Deyo RA. Imaging strategies for low-back pain: systematic review and meta-analysis. Lancet 2009;373(9662):463–472. https://doi.org/10.1016/S0140-6736(09)60172-0
- Jarvik JG et al. Rapid MRI vs radiographs for patients with low back pain: an RCT. JAMA 2003;289(21):2810. https://doi.org/10.1001/jama.289.21.2810
- Modic MT et al. Acute low back pain and radiculopathy: MR imaging findings and their prognostic role and effect on outcome. Radiology 2005;237(2). https://doi.org/10.1148/radiol.2372041509
- Lemmers GPG et al. Imaging versus no imaging for low back pain: a systematic review. Eur Spine J 2019. https://doi.org/10.1007/s00586-019-05918-1
- Hall AM, Aubrey-Bassler K, Thorne B, Maher CG. Do not routinely offer imaging for uncomplicated low back pain. BMJ 2021;372:n291. https://doi.org/10.1136/bmj.n291
- Hilton B et al. Route to diagnosis of degenerative cervical myelopathy in a UK healthcare system. BMJ Open 2019;9(5):e027000. https://doi.org/10.1136/bmjopen-2018-027000
- Gilbert FJ et al. Low back pain: influence of early MR imaging or CT on treatment and outcome. Radiology 2004;231(2). https://doi.org/10.1148/radiol.2312030886
- Weisman A, Masharawi Y. When imaging is normal, but biology is not. J Pain 2026. https://doi.org/10.1016/j.jpain.2026.106399
- Guo M et al. MRI-based radiomics reveals disc-level radiomic heterogeneity and is associated with symptom severity in single-level lumbar disc herniation. Eur Spine J 2026. https://doi.org/10.1007/s00586-026-10126-9