Shoulder Pain Not Improving After Physiotherapy: What Next?

Short answer: Shoulder pain that has not improved after physiotherapy does not automatically mean that rehabilitation cannot help or that surgery is required. The working diagnosis may be incomplete, the exercise dose may be too high or too low, the programme may not have progressed towards your actual goals, or another condition may need investigation. The next step is a structured reassessment—not simply repeating the same exercises indefinitely.

“I already tried physiotherapy” can describe very different situations: three sessions of passive treatment, six weeks of well-progressed exercise, a generic online programme, or months of rehabilitation after a significant injury. Before deciding that treatment has failed, it is important to define what was tried, whether it matched the suspected condition, and what changed in pain, sleep, movement, strength and function.

When persistent shoulder pain needs urgent assessment

Seek urgent medical assessment if:

• the shoulder is deformed or appears out of position after trauma;
• you suddenly cannot lift the arm after a fall, forceful pull or heavy lift, particularly with a snap, bruising or marked weakness;
• the shoulder is hot, red or rapidly swelling, especially with fever or feeling seriously unwell;
• the arm or hand becomes cold, pale or blue, or develops sudden substantial swelling;
• there is new or rapidly worsening weakness, extensive numbness, loss of coordination or difficulty walking;
• shoulder or arm discomfort occurs with chest pressure, breathlessness, sweating, nausea, faintness, or pain spreading to the jaw or back;
• pain is severe and rapidly worsening, or occurs with unexplained weight loss, an enlarging mass or a history of cancer.

Do not continue exercising through these features while waiting to see whether they settle. For a life-threatening emergency in Singapore, call SCDF 995.^[1]

Has physiotherapy truly failed?

Pain intensity is only one outcome. Early progress may appear as:

• less night disturbance;
• fewer painful episodes or faster recovery after activity;
• improved movement or strength;
• greater confidence using the arm;
• ability to work, dress, drive or exercise for longer;
• less reliance on medication;
• better tolerance of a previously painful task.

If none of these has changed, reassessment is appropriate. If function is improving while pain fluctuates, the plan may be working but still need refinement. A clear baseline makes that distinction easier than relying on memory alone.

1. The diagnosis may be wrong or incomplete

“Shoulder pain” is not a diagnosis. Similar symptoms may arise from:

• rotator-cuff-related shoulder pain or bursitis;
• a partial or full-thickness rotator cuff tear;
• frozen shoulder;
• glenohumeral or acromioclavicular-joint arthritis;
• calcific tendinopathy;
• biceps or labral pathology;
• shoulder instability;
• a fracture or bone stress problem;
• pain referred from the neck or irritation of a nerve;
• less common inflammatory, infectious or systemic disease.

A programme designed for a load-sensitive rotator cuff may not suit an irritable frozen shoulder. Exercises aimed only at the shoulder may not resolve nerve-related symptoms from the neck. New traumatic weakness needs a different evaluation from gradual soreness after increasing gym volume.

2. The initial condition may have changed

A reasonable diagnosis at the first visit may need updating. Progressive loss of passive movement can make frozen shoulder more apparent over time. New weakness after another injury may raise concern for a tear. Increasing rest pain, fever or systemic symptoms require a new safety screen.

Clinical reasoning is an ongoing process. Failure to follow the expected course is itself information and should trigger reconsideration rather than automatic escalation of the same treatment.

3. The programme may be too generic

A standard list of band exercises cannot account for every diagnosis, pain level or goal. A useful programme should answer:

• Which movement or capacity is currently limited?
• What starting dose can the shoulder tolerate?
• How will resistance, range, speed or repetitions progress?
• Which work, sport or daily task is being rebuilt?
• What response is acceptable during the exercise and over the next 24 hours?
• When should the plan be modified or the diagnosis reconsidered?

The 2025 rotator cuff tendinopathy clinical practice guideline supports an active rehabilitation programme as initial care, with exercises that may include motor-control and resistance training.^[2] “Active” does not mean identical exercises for everyone; selection and progression still require individualisation.

4. The exercise dose may be too high

More exercise is not always more effective. A large jump in load, repeated training to fatigue, forceful end-range stretching or insufficient recovery may keep an irritable shoulder flared. This is particularly relevant when pain disturbs sleep or remains substantially worse the following day.

The answer is not necessarily to stop all movement. It may be to reduce resistance, range, speed, repetitions or frequency, then rebuild from a tolerable level. In a painful frozen shoulder, aggressive stretching may be less acceptable than gentle movement matched to irritability.

5. The exercise dose may be too low

A programme can also remain below the stimulus needed to change strength or capacity. The same light band and repetitions may become comfortable but never prepare the shoulder for lifting luggage, serving in tennis, swimming or repeated overhead work.

Progressive loading means gradually changing demand as capacity improves. Progression may involve load, range, lever length, repetition, speed, coordination or task complexity. It should not be reduced to continually adding weight regardless of symptoms.

6. Treatment may have been mostly passive

Manual therapy, massage, heat, cold, taping and selected physical modalities may provide short-term symptom relief. That can be useful when pain prevents sleep or movement. However, passive care alone does not automatically restore strength, endurance, control or confidence for demanding tasks.

A passive treatment should have a defined purpose and review point. If it makes the shoulder temporarily more comfortable, the next question is whether that window is being used to restore movement and function.

7. The programme may not match your real-life load

Clinic exercises can improve while the patient’s main task remains unchanged. A desk worker, caregiver, warehouse worker, swimmer and badminton player expose the shoulder to different positions, loads, speeds and repetition.

Task-specific rehabilitation may need to reproduce the relevant height, weight and duration in stages. Work modification may be required while capacity develops. Without this bridge, a patient may feel stronger during exercises but still flare every time normal duties resume.

8. Adherence may be difficult for understandable reasons

A programme cannot help if it is impractical. Barriers may include pain, fear of damage, long work hours, unclear instructions, too many exercises, lack of equipment or uncertainty about how much discomfort is acceptable.

This is not a moral failure. The plan should be redesigned around what the patient can realistically do. Two or three clearly prioritised exercises may be more useful than a long routine that is rarely completed.

9. Sleep, recovery and health conditions may be influencing progress

Poor sleep can increase pain sensitivity and reduce recovery. Diabetes, thyroid disease and other health factors may be relevant in frozen shoulder. Smoking, medication, nutrition, stress and competing physical loads may also influence rehabilitation.

A biopsychosocial assessment considers these contributors without implying that shoulder pain is imaginary. Tissue health, nervous-system sensitivity, beliefs, sleep, work pressure and social demands can all affect the lived experience and the ability to follow a plan.

10. Expectations or timeframes may not match the condition

A mild load-related flare and a frozen shoulder do not recover on the same schedule. Tendon capacity and strength usually take time to build. Frozen shoulder can last many months. A significant traumatic injury may require protection or referral before progressive rehabilitation.

There should still be review points. “It takes time” should not become a reason to repeat unchanged treatment without measuring progress, especially when symptoms are worsening or function is declining.

What should a reassessment include?

A structured review may examine:

• the original onset, including trauma and changes in load;
• pain location, night symptoms and the exact aggravating tasks;
• active and passive shoulder movement;
• rotator cuff and broader upper-limb strength;
• stiffness, apprehension, clicking or instability;
• neck movement, sensation, reflexes and neurological strength where relevant;
• the exercises and treatments already tried, including dose and progression;
• adherence, barriers and the response over the following 24 hours;
• changes in sleep, work, sport and daily function;
• warning features and relevant medical conditions.

No single clinical test confirms every shoulder diagnosis. Findings are interpreted as a pattern. For example, loss of both active and passive movement suggests a different problem from near-normal passive movement with marked weakness after trauma.

When may an X-ray be considered?

X-ray may be appropriate after trauma or when arthritis, fracture, dislocation, calcific tendinopathy or another bone or joint condition could explain the lack of progress. For chronic shoulder pain, the American College of Radiology rates shoulder radiography as usually appropriate initial imaging.^[3]

An X-ray does not directly show the rotator cuff or labrum. A normal result therefore does not rule out all clinically important causes of shoulder pain.

When may ultrasound be considered?

Ultrasound can assess the rotator cuff, biceps tendon, bursa and calcific deposits. It can be useful when a focused tendon or bursal question would change management and can guide selected procedures.

It is operator dependent and does not provide the same overview of the labrum, bone marrow and deeper joint structures as MRI. The imaging choice should follow the clinical question rather than convenience alone.

When may MRI be considered?

MRI may be considered when there is substantial weakness after trauma, persistent symptoms despite an appropriate programme, suspected rotator cuff or labral injury, instability, unexplained stiffness, or a treatment or referral decision that requires more anatomical detail.

The ACR recommends selecting ultrasound, MRI or other imaging after initial radiographs according to the suspected source and existing findings.^[3] MRI uses magnetic fields rather than ionising radiation, but it is not automatically necessary because physiotherapy has not helped.

AAOS notes that MRI or ultrasound can demonstrate the location and size of a rotator cuff tear, while X-rays do not directly show the tear.^[4] Imaging findings still need clinical correlation because some tendon tears and age-related changes do not cause symptoms.

What treatment options may come next?

Revise the rehabilitation plan

A revised plan may change the working diagnosis, exercise selection, dose, progression or goal. It may temporarily reduce irritability before rebuilding movement, strength and task tolerance. The plan should specify what improvement is expected and when it will be reviewed.

Medication for a defined purpose

Appropriate medication may help reduce pain enough to sleep, work or participate in rehabilitation. Suitability depends on allergies, pregnancy, other medicines and medical conditions affecting the stomach, kidneys, heart, liver or bleeding risk. Medication should not obscure a worsening injury or replace reassessment.

Injection or another procedure

An injection may be considered for selected diagnoses after discussion of expected benefit, limitations, risks and alternatives. Its role may be to create a period in which movement and rehabilitation are more tolerable. Repeated procedures without a clear diagnosis, functional plan and review point should be questioned.

Referral for another opinion

Referral may be appropriate for a traumatic full-thickness tear, recurrent instability, advanced arthritis, fracture, neurological loss or another condition requiring a different pathway. Persistent pain alone does not mean surgery is inevitable.

Patient-specific scenarios

If you completed only a few physiotherapy sessions

That may have been enough for assessment and initial education but not enough to judge a progressive rehabilitation programme. Review whether you were given a clear diagnosis, home plan, progression criteria and realistic timeframe before concluding that physiotherapy cannot help.

If you completed six to twelve weeks with no measurable change

Check whether the programme was adhered to and appropriately progressed. If pain, function, movement and strength are all unchanged—or symptoms are worse—diagnostic reassessment and selected imaging may be reasonable rather than continuing unchanged care.

If every exercise makes the shoulder worse

Clarify whether the response is brief and tolerable or severe and prolonged. The diagnosis, irritability and exercise dose may need review. Progressive stiffness, traumatic weakness or neurological symptoms require particular attention.

If you prefer passive treatment

Discuss what made active care unacceptable. A selected passive modality, medication or procedure may help short-term symptoms, but the limitations should be explicit. A small programme linked to essential tasks may be more manageable than a lengthy gym routine.

If you need to return to work within one week

Define the required reach, load, repetition, driving and overhead duration. A short-term plan may combine symptom control, temporary duties and safe load modification while reassessment proceeds. No intervention can guarantee full recovery within a week.

If your MRI is normal but pain continues

A normal MRI can be reassuring about some structural concerns but does not invalidate pain. Load sensitivity, movement capacity, referred pain, nervous-system sensitivity, sleep and work demands may still be relevant. Revisit the clinical pattern rather than repeatedly searching for a single damaged structure.

If your MRI shows a tear

The significance depends on trauma, tear size and location, weakness, age, function and goals. Some tears can be managed nonsurgically; selected acute or functionally important tears may merit surgical assessment. The scan should inform shared decision-making rather than dictate it alone.

Integrated doctor and physiotherapy management

The Pain Relief Clinic has in-house physiotherapists for integrated management with the licensed medical doctor. This can be particularly useful when progress has stalled: the team can review the working diagnosis, compare examination findings with the rehabilitation response, adjust progressive loading, and coordinate decisions about imaging, medication, selected procedures or referral. Physiotherapists practising in Singapore are regulated by the Allied Health Professions Council, which maintains a public register.^[5]

The clinic can assist with clinically appropriate referrals and supporting paperwork for insurance claims. Coverage, required documents, pre-authorisation and reimbursement depend on the patient’s policy and insurer. A referral, form or medical report does not guarantee approval or payment.

What information should you bring to the reassessment?

• Previous scan reports and images if available.
• A list or video of the exercises prescribed.
• The resistance, repetitions and weekly frequency actually completed.
• A brief record of pain, sleep and next-day response.
• The work, sport or daily task that remains limited.
• A current medication list and relevant medical history.

This information helps distinguish an ineffective treatment from an incomplete dose, a poorly tolerated dose or a diagnosis that needs revision.

Frequently asked questions

How long should physiotherapy take to help shoulder pain?

There is no universal deadline. The expected course depends on the diagnosis, severity, duration, health factors and required function. A plan should nevertheless include measurable goals and review points rather than continuing indefinitely without change.

Does failed physiotherapy mean I need surgery?

No. The first question is whether the diagnosis and programme were appropriate. Revised rehabilitation, load modification, medication, selected procedures or further investigation may be considered before surgery, depending on the condition.

Should I stop my exercises if they hurt?

Not every discomfort is harmful, but severe pain, prolonged flares or worsening function should prompt adjustment. Stop and seek assessment for urgent warning features or sudden traumatic weakness.

Can I ask for imaging after physiotherapy has not helped?

Yes. Discuss which diagnosis remains uncertain and how the result would change care. X-ray, ultrasound and MRI answer different questions; the appropriate test depends on the clinical pattern.

Would changing physiotherapists help?

A second opinion may help if the diagnosis, goals, progression or communication remain unclear. Continuity can also be valuable when progress is being measured. The decision should focus on whether the current plan is specific, collaborative and responsive to change.

Can an injection make physiotherapy work better?

For selected diagnoses, short-term pain reduction may make movement and rehabilitation easier. Results and risks vary, and an injection does not replace progressive restoration of function.

Key points

• No improvement after physiotherapy should trigger reassessment, not endless repetition of the same plan.
• The diagnosis, exercise dose, progression, adherence and real-life task demands all matter.
• Passive treatments may support symptom control but do not automatically rebuild shoulder capacity.
• X-ray, ultrasound and MRI have different roles and should answer a defined clinical question.
• A scan finding does not determine treatment without the history, examination, function and patient priorities.
• Persistent shoulder pain does not automatically mean surgery is required.

Arrange a shoulder-pain reassessment in Singapore

If shoulder pain has not improved despite a properly followed programme, or if weakness, stiffness, night pain or loss of function is progressing, a structured reassessment can help decide whether the diagnosis, rehabilitation dose, imaging or treatment pathway should change. Seek urgent care for the warning features described above.

The Pain Relief Clinic
350 Orchard Road, #10-00 Shaw House, Singapore 238868
Call or WhatsApp: +65 9068 9605

Not every investigation or treatment discussed on this page is required, offered or recommended by the clinic. Availability and permissible use of tests, medicines, devices, injections and emerging treatments in Singapore may change with evolving Ministry of Health, Health Sciences Authority and Singapore Medical Council requirements. Individual decisions require clinical assessment and should follow current Singapore regulations and professional guidance.

References

1. Singapore Civil Defence Force. Emergency Medical Services.
2. Desmeules F, Roy JS, Lafrance S, et al. Rotator Cuff Tendinopathy Diagnosis, Nonsurgical Medical Care, and Rehabilitation: A Clinical Practice Guideline. J Orthop Sports Phys Ther. 2025;55(4):235–274.
3. American College of Radiology. ACR Appropriateness Criteria: Chronic Shoulder Pain.
4. American Academy of Orthopaedic Surgeons. Rotator Cuff Tears.
5. Allied Health Professions Council Singapore. Registers of Allied Health Professionals.

Disclaimer: This information is provided for general reference only and should be used in consultation with your licensed medical doctor. It is not a diagnosis or a substitute for personalised medical advice. Seek urgent care for severe, rapidly worsening or warning-feature symptoms.

Reviewed by Dr Terence Tan — 13 September 2026