Shoulder Pain: When Do You Need a Scan, Injection or Orthopaedic Opinion?
Shoulder pain can be frustrating, particularly when it persists despite treatment.
It is also an area where scans can sometimes create more qu
estions than answers. An MRI may identify tendinopathy, bursitis, degeneration or a rotator cuff tear, but does finding one of these abnormalities tell us what is causing the pain?
Not necessarily.
Equally, persistent shoulder pain does not automatically mean you need an injection or an orthopaedic opinion.
The important starting point is establishing the most likely diagnosis through a thorough clinical assessment. From there, we can decide whether rehabilitation needs to be optimised, whether imaging would genuinely add useful information, whether an injection has a role or whether there is actually a surgical question that needs answering.
Table of Contents
What can cause shoulder pain?
There is no single diagnosis that explains all shoulder pain.
Common presentations include rotator cuff related shoulder pain, frozen shoulder, osteoarthritis and problems involving the acromioclavicular joint.
Pain can also sometimes be referred to the shoulder from elsewhere, particularly the neck.
The history of the problem is therefore important.
Did the pain begin gradually or following an injury? Is movement restricted? Is there genuine weakness? Which movements are painful? Is the pain disturbing sleep? Are there symptoms travelling down the arm? What activities are you struggling to perform?
The answers help establish which diagnosis is most likely and whether further investigation is required.
This matters because treatment should be based on the clinical presentation rather than simply applying the same set of shoulder exercises to everyone with shoulder pain.
Do you need a scan for shoulder pain?
Not always.
Many shoulder problems can initially be assessed and managed without imaging.
There are circumstances where an X ray, ultrasound or MRI can provide extremely useful information. The important question is whether that information is actually needed to make a diagnosis or change what happens next.
This distinction has become particularly important following research examining what shoulder MRI scans show in people from the general population.
How common are rotator cuff abnormalities in people without shoulder pain?
Very common.
The FIMAGE study, published by Ibounig and colleagues in 2026, investigated shoulder MRI findings in 602 adults aged between 41 and 76.
The results are striking.
MRI finding | Prevalence |
Any rotator cuff abnormality | 98.7% |
Rotator cuff tendinopathy | 25% |
Partial thickness rotator cuff tear | 62% |
Full thickness rotator cuff tear | 11% |
Perhaps even more importantly, rotator cuff abnormalities were found in approximately 96% of asymptomatic shoulders and 98% of symptomatic shoulders.
Most of the full thickness tears identified were also found in shoulders without symptoms.
These findings do not mean that rotator cuff abnormalities are never responsible for shoulder pain.
They absolutely can be.
What they demonstrate is that identifying an abnormality on an MRI does not automatically tell us whether that abnormality is responsible for the symptoms.
The scan needs clinical context.
Does a rotator cuff tear on MRI mean it is causing your pain?
No.
This is where the language surrounding imaging becomes important.
Imagine having shoulder pain and being told that your MRI shows a tear.
It is entirely understandable to interpret that as something being damaged that needs repairing.
For some people, particularly following significant trauma associated with loss of function or weakness, a rotator cuff tear may be highly relevant and an orthopaedic opinion may be appropriate.
But for someone who develops shoulder pain gradually without a significant injury, the interpretation can be very different.
The FIMAGE study demonstrates why.
If partial thickness tears and other rotator cuff abnormalities are extremely common in people without pain, finding one in somebody who happens to have shoulder pain does not prove that it is the sole cause of their symptoms.
Age matters too.
Structural changes become increasingly common as we get older. Describing every change as damage or failure can therefore create an unnecessarily worrying picture of what may partly represent normal age related change.
The MRI report itself may be technically accurate.
The important question is what those findings mean for the individual.
Can rotator cuff problems improve with physiotherapy?
Yes.
Many people with rotator cuff related shoulder pain can improve with appropriate conservative management.
Exercise is usually an important part of this, but rehabilitation involves considerably more than simply being given several exercises for the shoulder.
The programme needs to reflect the person's symptoms, strength, physical capacity and goals.
Load and dosage matter.
So does progression.
Someone trying to return to gardening may require something very different from someone wanting to return to tennis, swimming or heavy strength training.
There is also evidence comparing conservative and surgical management.
Cederqvist and colleagues studied people with rotator cuff disease who had initially undergone a period of rehabilitation. At two year follow up, surgery did not provide superior outcomes to non surgical treatment for the overall population studied.
This does not mean rotator cuff surgery is ineffective or never appropriate.
It means that finding rotator cuff pathology does not automatically mean that surgery will provide a better outcome than well managed conservative care.
There are circumstances where surgical assessment becomes much more relevant, including certain traumatic tears and presentations involving significant loss of function.
Again, the diagnosis, clinical presentation and individual circumstances matter.
When might a shoulder injection help?
Corticosteroid injections can have a role in managing some shoulder conditions.
The important question is what we are trying to achieve by using one.
For some people, reducing pain temporarily may make it easier to sleep, move the shoulder or engage more effectively with rehabilitation.
In other situations, the likely benefit may be limited.
The diagnosis matters here too.
An injection for a highly irritable frozen shoulder, for example, is being considered in a different clinical context from an injection for rotator cuff related shoulder pain.
An injection should therefore not simply be viewed as the next treatment when exercises have not worked.
Before considering one, it is useful to understand:
What is the likely diagnosis?
What is the injection expected to achieve?
What are the potential benefits and limitations?
Has appropriate conservative management already been explored?
What is the plan following the injection?
An injection can sometimes be a useful part of the management strategy.
It is rarely the whole strategy.
When should you get an orthopaedic opinion for shoulder pain?
An orthopaedic opinion becomes particularly useful when there is a genuine surgical question to answer.
Examples might include a significant traumatic shoulder injury with suspected structural damage, substantial weakness or loss of function associated with a clinically relevant rotator cuff tear, advanced symptomatic osteoarthritis, or persistent symptoms where appropriate conservative management has been explored and surgery may reasonably be considered.
There may also be other shoulder conditions where referral is appropriate depending on the clinical findings.
What I would avoid is assuming that persistent pain automatically means someone needs to progress from physiotherapy to an MRI and then to a consultant.
Healthcare does not need to operate as a ladder.
An orthopaedic surgeon has specialist expertise in determining whether surgical management is appropriate.
An experienced musculoskeletal clinician has specialist expertise in assessing the clinical presentation and optimising conservative management.
The appropriate clinician therefore depends on the question that needs answering.
What if your shoulder pain is not improving?
If your shoulder remains painful despite treatment, the first step should usually be to reassess the situation rather than simply repeat the same treatment.
Several questions become important.
Was the original diagnosis correct?
Was it clearly explained?
Has rehabilitation genuinely been individualised?
Was exercise load appropriate?
Was the programme progressed?
Was sufficient time allowed for the condition being treated?
Has anything changed since the original assessment?
Are there findings that now make imaging useful?
Would an injection help facilitate rehabilitation?
Is there now a genuine reason to obtain an orthopaedic opinion?
Sometimes conservative management has been appropriate but simply needs more time.
Sometimes the rehabilitation programme can be significantly improved.
Sometimes the diagnosis needs reconsidering.
And sometimes escalation is exactly the right decision.
How do you decide what the next step should be?
This is where clinical reasoning becomes particularly important.
There is no benefit in being automatically against scans, injections or surgery.
All three can be extremely useful when used appropriately.
Equally, there is no benefit in escalating through them simply because shoulder pain has persisted.
A thorough assessment should bring together the history, examination findings, previous treatment, functional limitations and any existing imaging.
If a scan has already been performed, the findings need to be clinically correlated.
A rotator cuff tear may be highly relevant.
It may also be an incidental finding.
The difference cannot always be established by reading the MRI report alone.
My previous role as an Advanced Physiotherapy Practitioner within the NHS involved assessing more complex musculoskeletal presentations and making these types of decisions around diagnosis, imaging, conservative management and onward referral.
At Genuine Physio in Tadworth, near Kingswood, the aim is therefore not to favour
physiotherapy over an injection or an orthopaedic opinion.
It is to understand the problem properly and help determine which option is most appropriate.
Sometimes that means optimising rehabilitation.
Sometimes imaging is needed.
Sometimes an injection is worth considering.
Sometimes an orthopaedic opinion is exactly the right next step.
Good musculoskeletal care is about knowing the difference.
Frequently Asked Questions
Do I need an MRI for shoulder pain?
Not necessarily. Many shoulder problems can initially be diagnosed and managed through clinical assessment. MRI becomes more useful when the findings are likely to clarify an uncertain diagnosis or influence management.
Does a rotator cuff tear always cause pain?
No. Rotator cuff tears and other abnormalities are commonly found on MRI in people without shoulder pain. Whether a tear is clinically relevant depends on the history, examination findings and individual presentation.
Can a rotator cuff tear heal without surgery?
Many people with rotator cuff tears can achieve meaningful improvements in pain and function without surgery. Whether surgery should be considered depends on factors including the type of tear, how it occurred, weakness, functional loss, symptoms and response to conservative management.
Is physiotherapy or a steroid injection better for shoulder pain?
It depends on the diagnosis and clinical presentation. Rehabilitation is important for many shoulder conditions. A corticosteroid injection may sometimes provide short term symptom relief and can occasionally help someone engage more effectively with rehabilitation. It should not automatically replace an appropriate rehabilitation programme.
When should I see an orthopaedic consultant for shoulder pain?
An orthopaedic opinion is particularly appropriate when there is a genuine surgical question, such as certain traumatic injuries, clinically significant structural pathology or persistent symptoms despite appropriately optimised conservative management.
What if my MRI says I have a rotator cuff tear?
Do not interpret the word “tear” in isolation. Rotator cuff tears are common, particularly as we get older, and can exist without pain. The MRI findings should be considered alongside your symptoms and clinical examination to determine whether the tear is actually relevant.
Unsure what to do about persistent shoulder pain?
If you have persistent shoulder pain and are unsure whether you need further rehabilitation, a scan, an injection or an orthopaedic opinion, get in touch with Genuine Physio.
A detailed assessment can help establish the most likely diagnosis, consider whether any imaging findings are genuinely relevant and determine the most appropriate next step.
References
Ibounig T, et al. FIMAGE study. Population based MRI assessment of rotator cuff abnormalities in adults aged 41 to 76 years. JAMA Internal Medicine. 2026.
Cederqvist S, et al. Non surgical and surgical treatments for rotator cuff disease. Two year outcomes of a randomised clinical trial. 2020.
Comments