Shoulder pain: what physiotherapy achieves – Knowledge

1. Why this article?

Shoulder pain is among the most common reasons for which people come to physiotherapy. A review of 61 studies from various countries arrives at an average of 16 per cent: that is how many adults have shoulder pain within a year [2]. In general practice, shoulder complaints rank third among musculoskeletal problems, after the back and the neck [1].

People who take that problem to a doctor or a physiotherapist often hear very different explanations: impingement, entrapment syndrome, tendon irritation, bursitis, a rotator cuff tear, calcification, poor posture, a weak shoulder blade. Each of these labels names a structure and thereby suggests that this particular structure is producing the pain. Research over the last fifteen years has shown that this attribution is far more uncertain than the terms sound. That has consequences for which investigations make sense, which treatment helps and what you may expect.

This article answers six questions:

• How is the shoulder built, and why is this joint in particular so vulnerable?
• Which forms of shoulder pain are distinguished, and how are they recognised?
• What does imaging achieve, and in which situations does it do more harm than good?
• What does physiotherapy deliver, measured against what studies show?
• What does an exercise programme look like: which exercises, how heavy, how often, how long?
• When are medication, injections or surgery appropriate?

The article is written for those affected and for their families. Technical terms appear because you will meet them in the medical report anyway. Each one is explained when it first comes up. The numbers in square brackets refer to the reference list at the end.

One note in advance: this article does not replace an examination. The shoulder is one of the few regions of the body into which complaints from the heart, the lungs and the cervical spine can also radiate. A careful assessment by a doctor or a physiotherapist therefore comes first, and what is decided there for your situation takes precedence over this text.

2. How the shoulder is built

To understand why the shoulder hurts so often and why training plays such a large role here, a brief look at the construction helps.

2.1 Four connections working together

What everyday language calls «the shoulder» consists of four connections that work together:

• The shoulder joint itself (technical term: glenohumeral joint, from «glenoid» for the socket and «humerus» for the upper arm bone). The round head of the upper arm bone rests on a small, almost flat socket on the shoulder blade. A common comparison is a golf ball on a tee: the mobility is enormous, the bony guidance correspondingly small.
• The sliding surface between shoulder blade and rib cage. The shoulder blade lies on the back of the rib cage and travels along when the arm is raised. About one third of the total arm elevation comes from this travel.
• The acromioclavicular joint (AC joint for short). It sits right on top of the shoulder, where the collarbone meets the roof of the shoulder. You can feel it with a finger as a small step.
• The joint between collarbone and breastbone at the front in the middle. It is the only bony connection of the whole arm to the trunk.

This construction explains the first peculiarity of the shoulder: it is the most mobile joint in the body, and its stability has to come predominantly from muscles and tendons.

2.2 The rotator cuff

Four muscles arise from the shoulder blade and run with their tendons around the head of the upper arm bone like a cuff. This group is called the rotator cuff:

• Supraspinatus runs above the joint and helps to lift the arm out to the side. Its tendon is the one most often involved in complaints.
• Infraspinatus and teres minor lie at the back and turn the arm outwards.
• Subscapularis lies at the front between shoulder blade and ribs and turns the arm inwards.

The actual task of these four muscles lies less in providing force for large movements. They centre the head of the upper arm bone in the flat socket while the large muscles – deltoid, pectoral, latissimus – move the arm. You can picture the cuff as a guiding crew that holds the head in place so that the powerful engines can pull cleanly.

Above the supraspinatus lies the subacromial space: the gap between the roof of the shoulder (technical term: acromion) and the head of the upper arm bone. In this gap sits a bursa (technical term: subacromial bursa), a flat gliding cushion that pads the tendon against the bone. In addition, the long tendon of the biceps muscle runs through a groove at the front of the upper arm bone into the joint.

2.3 What follows from this

Three points follow from this construction, and they carry the rest of this article:

• The shoulder depends on its muscles. Where ligaments and bone provide little guidance, muscles and tendons take over. This is an essential reason why targeted training ranks so high on the list for shoulder pain.
• The space under the roof of the shoulder is narrow. That is exactly where most people feel the pain, and exactly where conditions change with age, with load and with training status.
• Many structures lie close together. Tendon, bursa, joint capsule and biceps tendon are within a few centimetres of each other. It is therefore hardly possible from the outside to determine which of these structures produces the pain. This circumstance explains the next section.

3. The many names for the same problem

3.1 Why the terms have changed

For decades shoulder pain was named after the presumed structure: tendinitis, tendon degeneration, partial tear, bursitis, impingement syndrome. These names assume that the examination can reliably say which structure is affected. That, however, rarely succeeds.

Three observations have led to a rethink [1][3]:

• The terms were defined very differently across studies. What one paper called «impingement» was «rotator cuff tendinopathy» in the next. Comparing results becomes difficult as a result.
• The special tests of the physical examination agree only moderately between two clinicians and cannot reliably name the source of the pain.
• The findings on ultrasound and MRI are present in very many people who have no complaints at all. More on this in section 6.

An umbrella term has therefore taken hold that describes rather than asserts. The review in JAMA Internal Medicine of August 2026 recommends the expression subacromial pain [1]. In physiotherapy the expression rotator cuff–related shoulder pain is also in use [3]. Both terms say the same thing: the pain sits in the region under the roof of the shoulder and is connected with the tendons of the rotator cuff, and which of the structures lying there produces it remains open.

3.2 Why the choice of words has practical consequences

The name of a diagnosis influences what people think about their shoulder and which treatment they consider necessary. In a randomised experiment, respondents were presented with the same case history carrying different labels [4]. Terms such as «rotator cuff tear» led respondents to think more often of damage that has to be repaired, and to consider surgery or imaging necessary more often. Descriptive terms triggered this train of thought less often.

The choice of words thus has entirely practical consequences. Anyone who believes that something in their shoulder is torn and must first be stitched will be more cautious when exercising, will tend to spare the arm and will tend to avoid loading it. That very behaviour slows recovery. An understandable explanation of what is actually going on in the shoulder therefore belongs to the treatment and not merely in front of it.

4. Which forms are distinguished

Fine attribution within the subacromial region is difficult. Some forms of shoulder pain, by contrast, can be told apart reliably, and this distinction changes the treatment. It rests on the case history and on a few manoeuvres in the examination.

4.1 Subacromial pain

This is by far the most common form. In general practice it accounts for about 90 to 95 per cent of people with shoulder pain that did not follow an injury [1]. The typical picture is:

• The pain sits on the outside of the upper arm, roughly where the deltoid muscle attaches, that is a hand's width below the top of the shoulder.
• It appears when the arm is raised, especially above shoulder height, and when reaching behind the back.
• At night, lying on the affected side hurts. Disturbed sleep is the greatest burden for many of those affected.
• On examination, active movement is painful, often in a middle range of arm elevation. Passive mobility – when someone else moves the relaxed arm – is preserved.
• Those affected are predominantly over 40 years of age.

4.2 Frozen shoulder

Frozen shoulder (technical term: adhesive capsulitis) arises from inflammation and subsequent shrinking of the joint capsule. Two things are characteristic: very severe pain, including at rest and at night, and a restriction of movement that persists even when someone else guides the arm. External rotation in particular is restricted: the elbow stays at the body and the forearm can hardly be turned outwards.

Those affected are usually between 40 and 65 years old, women more often than men, and people with diabetes considerably more often than others [1]. In the early phase, frozen shoulder resembles subacromial pain so closely that a distinction is not yet possible. The restriction of passive mobility develops only over the course. More in section 15.

4.3 Osteoarthritis of the shoulder joint

Here the joint cartilage is altered. The pain sits deeper in the joint, appears at first under load and becomes more persistent over the course. As with frozen shoulder, passive mobility is restricted too; in addition, a grinding or crunching can often be felt. Those affected are predominantly over 60 years of age [1]. There is a separate article on osteoarthritis in general under Osteoarthritis.

4.4 Osteoarthritis of the acromioclavicular joint

With the AC joint the pain sits precisely on top of the shoulder, and those affected can point to it with one finger. It intensifies when the arm is brought across the body to the opposite side, and when lying on the shoulder. Frequently there was a fall onto the tip of the shoulder in the past. Those affected are often physically active people between 40 and 65 years of age [1].

4.5 The long head of the biceps tendon

The pain lies at the front of the shoulder, in the groove in which the tendon runs. It intensifies with overhead work and when bending the elbow against resistance. If this tendon ruptures, a visible bulge of muscle appears on the upper arm, which specialists call the Popeye sign [1]. This rupture as a rule affects an already heavily degenerated tendon, causes surprisingly little loss of strength and is usually treated without surgery.

4.6 Calcific tendinitis

Calcium is deposited in one of the rotator cuff tendons. The complaints often come in episodes, and an acute phase can be extremely painful. Those affected are usually between 30 and 60 years old. Calcium deposits are, however, also found in up to a quarter of pain-free shoulders on plain radiographs [1]. A deposit alone therefore does not yet explain pain. More on this in section 17.

4.7 Pain that arises outside the shoulder

Some complaints are felt in the shoulder and arise elsewhere. This point is the most important one in the whole section, because it changes the treatment path entirely [1]:

• Cervical spine. An irritated nerve in the neck produces pain radiating into shoulder, arm and hand, often accompanied by tingling, numbness or loss of strength. An important distinguishing feature: shoulder movement itself does not trigger the pain, and the shoulder remains freely mobile.
• Heart. A sense of pressure or tightness in the chest pulling into the left shoulder, the arm or the jaw, together with breathlessness, sweating or nausea, requires immediate medical assessment. Women report shoulder pain in this situation more often than men.
• Lung. A tumour at the apex of the lung can produce shoulder pain radiating towards the inner side of the arm. Persistent shoulder pain in people who smoke therefore deserves particular attention.
• Brachial plexus. In neuralgic amyotrophy (technical term: Parsonage-Turner syndrome), very severe pain sets in suddenly, followed after days by marked weakness. It often occurs after an infection or a vaccination.

The examination at the outset serves first and foremost to exclude these possibilities. Once that has been done, the further treatment is largely similar across the great majority of forms of shoulder pain [1].

5. What the physical examination can achieve

The examination of the shoulder has two tasks, and these two tasks can be fulfilled to very different degrees.

5.1 The most important manoeuvre: active against passive

The first task succeeds reliably: the broad classification. It rests on a simple comparison [1]:

• Active movement means that you move the arm yourself. Lifting out to the side, lifting to the front, external rotation with the elbow at the body and reaching behind the back are all examined, always in comparison with the healthy arm.
• Passive movement means that you let go and the clinician moves the arm.

The relationship between the two gives the direction:

• Painful or restricted actively, but freely mobile passively: this speaks for subacromial pain.
• Restricted actively and passively, above all in external rotation: this speaks for frozen shoulder or for osteoarthritis of the shoulder joint.
• Shoulder movement free and pain-free, yet complaints present: this directs attention to the neck or to the internal organs.

Strength is also tested, by having you press outwards and inwards against resistance, again in comparison with the other side. Marked weakness can come from pain, from a large tendon tear or from a nerve disorder. In addition there are the so-called lag signs: the clinician brings your arm into an end position and lets go. If you cannot hold this position, that speaks for an extensive tear.

5.2 Why the special tests say little

The second task succeeds poorly: attributing the pain to a particular tendon or to a bursa. There is a whole series of named tests for this, such as the Neer test, the Hawkins-Kennedy test, the Speed test or the empty can test. The review in JAMA Internal Medicine explicitly advises caution in interpreting these tests [1]. Three reasons are given:

• Two experienced clinicians frequently reach different results in the same person.
• The tests cannot reliably name the source of the pain, because in the narrow space under the roof of the shoulder every one of the structures moves along.
• The studies of accuracy come predominantly from specialist clinics, in which considerably more severe cases occur than in an ordinary practice. The figures therefore cannot simply be transferred.

This does not devalue the examination. It clarifies whether a warning sign is present, how severe the restriction is, which movements and which loads provoke the complaints and where your strength stands. These are precisely the details that determine the exercise programme. The name of an individual tendon changes little about it.

6. Imaging: what ultrasound and MRI show

6.1 The decisive figure

Between February 2023 and April 2024, a Finnish research group invited 602 people from the general population aged 41 to 76 and examined both shoulders of every person by MRI, regardless of whether they had complaints. The result [5]:

• 99 out of 100 people had at least one abnormality of the rotator cuff. In detail: 25 per cent a tendon change without a tear, 62 per cent a partial tear, 11 per cent a full-thickness tear.
• Among the pain-free shoulders, the proportion with abnormalities was 96 per cent (1039 of 1076 shoulders); among the painful ones it was 98 per cent (126 of 128 shoulders).
• Only full-thickness tears were more common in painful shoulders (14.6 against 6.5 per cent), and even this difference disappeared once other influences were taken into account.

The authors conclude from this that changes to the rotator cuff belong to the normal picture in adults of middle and older age, much like grey hair or declining eyesight. A systematic review of 53 studies reaches the same result: abnormal findings in pain-free shoulders are the rule, and their frequency rises with age [6].

From this follows the practically most important sentence of this section: an abnormal image does not prove that this particular finding is causing your pain. In a person of your age with a pain-free shoulder, the same thing would very probably have been seen.

6.2 When an image is nevertheless sensible

The recommendation is to refrain from imaging at first in newly arisen shoulder pain without an injury and without warning signs [1]. A randomised trial in people with shoulder pain of less than three months' duration found no difference in self-assessed recovery after one year between management tailored by ultrasound and usual care.

Imaging is recommended when one of the following situations applies [1]:

• Suspicion of a serious condition such as a joint infection or a tumour.
• A major injury in which a fracture or a dislocation is possible.
• Marked or increasing weakness pointing to an extensive tear or a nerve disorder.
• Complaints that worsen over time instead of improving or staying the same.

There is deliberately no fixed interval after which an image would fall due, because recovery times vary widely. If an image is taken, one begins with a plain radiograph; it shows advanced osteoarthritis, rare bone tumours and a high-riding humeral head in long-standing tendon damage. Ultrasound and MRI are similarly accurate for full-thickness tears; MRI additionally shows the state of the muscle and is therefore preferred when surgery is under discussion.

The reason for this restraint lies in the consequences of imaging, far less in its cost. A finding discovered by chance that has nothing to do with the pain creates worry, leads to further investigations and occasionally ends in a treatment that helps nobody [1].

7. Warning signs: when this is not a case for physiotherapy

The following situations belong promptly in medical hands. They are rare, and for that very reason it is worth knowing them [1]:

• Fever, redness, warmth and swelling together with very severe pain: suspicion of a joint infection. This is an emergency.
• Unintended weight loss, night sweats, a history of cancer.
• Visible deformity or marked loss of strength after an injury: suspicion of a fracture, a dislocation or a fresh tendon rupture.
• Chest pain, breathlessness, nausea, sweating: suspicion of a heart problem. Call the emergency number immediately.
• Increasing paralysis or numbness in the arm.
• Complaints on both sides at once, together with morning stiffness and complaints in further joints: suspicion of an inflammatory rheumatic disease such as polymyalgia rheumatica.

8. What the course looks like without treatment

This question is rarely asked and is decisive for judging any treatment. If complaints recede on their own, a treatment has to achieve more than merely being present while things improve.

The figures from general practice look like this:

• About 80 per cent of people with subacromial pain experience improvement within a year, frequently sooner [1].
• In a Dutch study of 526 adults who newly consulted their general practitioner for shoulder complaints and were followed for ten years, close to 80 per cent had improved satisfactorily by the end of that period, with treatment usually consisting of watchful waiting or an anti-inflammatory painkiller [7].
• On the other hand: at six months a considerable proportion still have complaints. In a further Dutch study, roughly half were not yet free of complaints after six months [8]. Recovery is thus the rule, but in many people it takes months.

What makes better recovery at six months more likely: a sudden onset, stronger pain at the start, less restriction in everyday life at the start, and less of a tendency to expect the worst [1].

Two conclusions follow. First, you may count on things getting better, and this information is part of the treatment. Second, the sentence «X helped me» is weak as evidence, because many people would have improved in the same period without X. For this reason the next section draws above all on studies in which a treatment was compared with a sham treatment or with watchful waiting.

9. What physiotherapy demonstrably achieves

9.1 Education and advice

A systematic review of 14 studies with 5655 participants concludes that education can improve people's knowledge, their assessment of the need for treatment and their expectation of recovery [9]. In substance it covers four points: that shoulder pain is common, that it rarely indicates anything dangerous, that it usually improves over time, and how to organise everyday life in the meantime.

Organising everyday life includes: adapting demanding activities temporarily rather than immobilising the arm; alternating load and recovery (technical term: pacing); using painkillers as needed to make movement possible; and resuming activities as soon as they are manageable.

9.2 Exercise

The meta-analysis by Steuri and colleagues summarised the randomised trials of all non-surgical treatments for subacromial pain [11]. For pain, the following comparisons emerged among others:

• Exercise against no exercise: a clear advantage for exercise (standardised mean difference −0.94).
• Specific against general exercise: a moderate advantage for the specific programmes (−0.65).
• Manual therapy in addition to exercise against exercise alone: a small advantage, and only at the earliest follow-up (−0.32).

To place these figures: the standardised mean difference states by how many spreads two groups differ. Values around 0.2 count as small, around 0.5 as moderate, from 0.8 as large. A negative sign means less pain.

A recent review of 28 studies with 1702 affected people specifically examined whether strength can be improved [13]. The result is unambiguous: programmes with active exercises or with strength training improved strength in all directions, most clearly external rotation (standardised mean difference 0.56). Programmes without an exercise component did not improve strength at all.

9.3 The uncomfortable trial

The largest investigation on this topic is the GRASP trial with 708 participants from the United Kingdom [10]. All had had shoulder pain attributed to a rotator cuff disorder for at most six months. They were allocated to four groups, combined from two questions:

• A progressive exercise programme with up to six supervised sessions against a single advice session with a physiotherapist, in which the person received best-practice advice, simple home exercises and guidance on progressing independently.
• A corticosteroid injection under the roof of the shoulder at the start against no injection.

After twelve months the groups differed practically not at all on the primary outcome, a questionnaire on pain and function: the supervised exercise programme was 0.66 points ahead of advice (99 per cent confidence interval −4.52 to 3.20), the injection 1.11 points ahead of no injection (−4.47 to 2.26). At eight weeks the injection showed a small advantage of 5.64 points, most clearly in those with the highest pain scores at the start.

9.4 How this fits together

At first sight these results contradict each other. On closer inspection they fit together, and the resolution is the most important part of this article for practice.

First: the comparison standard in GRASP was high. The comparison group received a thorough consultation with a physiotherapist, a home exercise programme and guidance on how to progress independently. That is a treatment containing the greater part of what physiotherapy consists of. The trial therefore shows that six supervised sessions add little after a year compared with one well-conducted session plus a home programme. It does not show that movement is without effect.

Second: the strength of the evidence is limited. The meta-analysis by Steuri rates the certainty of its own statements as «very low» [11]. The Cochrane review on manual therapy and exercise found only small, clinically unimportant short-term benefits for a supervised exercise programme compared with a placebo [18]. The review by Naunton and colleagues concluded that progressive exercise with resistance may have a benefit but that the evidence for it remains uncertain, while exercise without resistance and without progression showed no benefit [15].

Third: the benefit shows where it is measured. For strength the evidence is unambiguous [13]. For pain after one year it is not, because by then most people have improved anyway. For everyday life both matter: someone who can put things away in the cupboard again after four months has gained something, even if at twelve months they would be where they would also have been without exercise.

From this follows the stance taken by the current clinical practice guideline on rotator cuff tendinopathy [12]: education, an exercise programme tailored to the person, regular review of progress, and return to activity guided by what works rather than by what the image shows. Exercise is thereby an offer with low risk, manageable effort and a benefit that is better established for strength and confidence than for the long-term course of pain.

10. What an exercise programme looks like

10.1 The building blocks

A programme for subacromial pain usually consists of four parts. The selection depends on which movements and which loads cause complaints in your case.

• Maintaining mobility. Swinging the relaxed arm, guided lifting with the healthy arm or with a stick, gentle stretching of the back of the shoulder capsule. These exercises keep the joint moving while loading still hurts.
• Strength of the rotator cuff. External and internal rotation against an elastic band or with a small dumbbell, starting with the elbow at the body. External rotation responded most clearly in the studies [13].
• Strength of the muscles that guide the shoulder blade. Rowing, pulling towards the body, lifting the arm in the diagonal plane. These exercises build up the muscles that hold the shoulder blade to the rib cage.
• Loading in everyday patterns. Lifting, carrying, supporting, working above shoulder height. This part comes later and depends on your occupation and your sport.

Two studies have examined whether the choice of exercises matters. The meta-analysis by Steuri found an advantage of specific over general exercise [11]. A Swedish randomised trial compared a specific programme emphasising external rotation and the shoulder blade muscles against a non-specific programme in people already on the waiting list for surgery. After three months the specific group was considerably better, and far fewer of these people still wanted to be operated on [16]. A more recent review that broke exercise programmes down by frequency, intensity, type and time found, however, only small differences between most variants [14].

The reasonable summary of these findings is: a programme that deliberately includes external rotation and the shoulder blade muscles and that progresses systematically is preferable to an arbitrary collection of shoulder exercises. Within that specification the precise choice of exercises is secondary, and that leaves room to pick exercises you are actually willing to do.

10.2 Dose

The programmes that worked in the studies looked roughly like this:

• Frequency: daily up to three times a week. A daily programme has the advantage of becoming a habit; three sessions a week are enough when the load is higher.
• Volume: two to three sets per exercise with eight to fifteen repetitions, three to five exercises, in total between ten and thirty minutes.
• Duration: at least eight to twelve weeks before an assessment makes sense.

The question of whether heavier weights are better than lighter ones has not yet been answered. The review organised by frequency, intensity, type and time found no reliable difference between high and low load [14]. What is emerging, by contrast: programmes that increase the resistance over the weeks do better than programmes that stay at the same level [15]. There is a separate article on the basics of strength training under Strength training.

10.3 Is exercise allowed to hurt?

Almost everyone affected asks this question, and it has a clear answer. A meta-analysis of studies in people with persistent musculoskeletal pain compared exercise programmes that deliberately allow temporary pain with programmes that stay strictly in the pain-free range [17]. In the short term the programmes with permitted pain had a slight advantage; in the medium and long term the two were equivalent. The authors conclude that pain during exercise does not stand in the way of success.

In practice the following rule of thumb has proved useful:

• Pain during the exercise up to about 4 or 5 on a scale of 0 to 10 is acceptable.
• The pain should return to the starting level within about 24 hours.
• If you have considerably more discomfort the next morning than the day before, the load was too high. Next time you work with less weight, fewer repetitions or a smaller range of movement.

This rule of thumb applies to the common forms of shoulder pain without warning signs. It does not apply to a very painful frozen shoulder in the early phase; there the load is reduced according to the pain (section 15).

10.4 Progressing

Progress comes from the increase in load, and there is an order for it that has proved itself:

1. Range of movement. First the range in which you exercise is enlarged, for instance from hip height up to shoulder height.
2. Repetitions. Then more repetitions are added, for instance from eight to fifteen.
3. Resistance. Then the band becomes stronger or the dumbbell heavier, and the number of repetitions goes back down.
4. Lever and speed. Finally the arm is taken further from the body and the movements are performed faster, because everyday life demands both.

A practical rule: if you manage the top number of repetitions cleanly on two training days in a row without an increase in complaints, raise the resistance next time.

10.5 How long

The studies measured predominantly over six to twelve weeks. To judge whether a programme is working for you, eight to twelve weeks is an appropriate period. For maintaining what has been achieved, the same applies as for any training: what is no longer loaded declines. Two sessions a week are enough to maintain strength, and these two sessions may be built into general strength training.

11. Manual therapy, devices and tape

Alongside exercise, physiotherapy offers a range of approaches in which the clinician does something to you rather than you doing something. The evidence for them, in order:

• Manual therapy (mobilisation of the joint, manipulation, massage). Against a sham treatment, the meta-analysis by Steuri showed a small advantage for pain (standardised mean difference −0.35) [11]. Given in addition to exercise, it was superior to exercise alone, but only at the earliest follow-up (−0.32). The Cochrane review classifies the benefit over a sham treatment as small to absent [18]. A defensible position: manual therapy can provide short-term relief and thereby make exercise possible. As a sole treatment over weeks it has no basis.
• Therapeutic ultrasound. The Cochrane review on device-based treatments found small short-term improvements in pain at low certainty of evidence, and mainly in the presence of calcium deposits [19].
• Laser. A similar picture: short-term improvements, essentially with calcium deposits, at low certainty [19]. In the meta-analysis by Steuri, laser was superior to a sham treatment (−0.88), and laser in addition to exercise was superior to exercise alone (−0.65) [11].
• Extracorporeal shockwave therapy. The Cochrane review summarises that shockwave provides little to no benefit over a sham treatment, at moderate certainty of evidence [20]. The review in JAMA Internal Medicine advises against it for subacromial pain [1].
• Tape. In the meta-analysis by Steuri, tape was superior to a sham application (−0.64), based on five studies with 272 people [11]. The effect is small to moderate, the risk low, and the benefit probably lies in the short term.

These approaches have a sensible role as door openers: if a treatment provides enough relief for a week that you can start on the exercises, it has served its purpose. Building strength and load tolerance remains the task of the exercises.

12. Posture, the shoulder blade and other persistent explanations

Two explanations are especially persistent, and both deserve a close look.

Posture. The idea that a rounded back or forward-hanging shoulders cause shoulder pain is widespread. Investigations of this show an inconsistent picture: on average, people with and without complaints hardly differ in posture, and the connection between postural features and later complaints is weak. A systematic review of shoulder blade position concluded that findings in people with and without subacromial pain overlap and that a uniform malposition cannot be demonstrated [21].

What follows: strengthening the muscles that guide the shoulder blade remains sensible, because it raises load tolerance. The justification lies in building strength and not in correcting a malposition. For everyday life this also means that you need not worry constantly about your posture. Changing position over the course of the day achieves more than the effort to hold one particular position.

The narrowness under the roof of the shoulder. The term «impingement» rests on the idea that the tendon is pinched between the head of the upper arm bone and the acromion when the arm is raised, and that surgery to widen the space must therefore help. This idea was refuted by the surgical trials described in section 14. Today's explanation assumes an interplay: the load tolerance of the tissue, the load applied, inflammatory and age-related changes in the tendon, and the processing of pain signals in the nervous system [1]. This explanation is less vivid than the image of pinching and fits better with what treatments actually accomplish.

13. Medication and injections

Medication has a modest and clearly delimited aim in shoulder pain: enough relief for movement and sleep to become possible. The evidence in detail [1]:

• Paracetamol. There are no direct studies for the shoulder. Investigations in other musculoskeletal regions indicate little to no benefit.
• Topical anti-inflammatory painkillers (gel or patch with a non-steroidal anti-inflammatory drug). A sensible first choice, provided the skin is intact and there is no allergy. In acute musculoskeletal complaints, pain relief was comparable to that of tablets, with less exposure for the body as a whole.
• Oral anti-inflammatory painkillers. Defensible as needed and for a short period. The benefit is small, and the risks for stomach, kidneys and circulation have to be weighed, especially in older people and in kidney disease.
• Opioids. Not recommended for shoulder pain. The expected benefit is small and the risk of side effects and dependence is high.

The corticosteroid injection under the roof of the shoulder is the most effective short-term measure and at the same time the one most often overestimated. A meta-analysis of randomised trials describes the benefit as small and transient: the advantage over the comparison treatment shows in the first weeks and is no longer detectable after about three months [22]. The GRASP trial found a small advantage at eight weeks and none at twelve months [10].

From this follows when an injection makes sense: with severe pain, especially night pain that prevents sleep, and with the explicit purpose of opening a window of time for the exercises. Further points that belong in the conversation [1]:

• Ultrasound guidance does not improve the result compared with an injection guided by landmarks.
• People with diabetes should expect a temporary rise in blood glucose.
• Caution is advised before and after shoulder surgery, because injections can raise the risk of infection and impair tissue healing.
• Hyaluronic acid and platelet-rich plasma are not recommended for subacromial pain, because a meaningful benefit over a sham treatment is lacking.

14. Surgery

On no other treatment of the shoulder has such thorough clarity been created over the last fifteen years.

Subacromial decompression is an arthroscopic procedure in which bone is removed from the acromion and the bursa is taken out, in order to create more room in the space under the roof of the shoulder. The Cochrane review of this procedure summarises two trials with 284 participants in which real operations were compared with placebo operations: after twelve months the difference in pain was 0.3 points on a scale of 0 to 10 (confidence interval −0.3 to 0.8), at high certainty of evidence [23]. For comparison: a difference of about 1.5 points counts as noticeable.

The Finnish FIMPACT trial, one of these two investigations, followed its participants for ten years and published the long-term results in December 2025: even after ten years the group with the real operation did not differ from the group with the placebo operation [24]. An international guideline panel derived a clear recommendation against this procedure for subacromial pain as early as 2019 [25], and the 2026 review confirms it [1]. The procedure also carries a risk: about 6 in 1000 people who undergo it suffer a serious complication [23].

The precise scope of this statement matters. It concerns subacromial pain without a full-thickness tendon tear. For tears, for frozen shoulder and for advanced osteoarthritis, separate considerations apply, and they are set out in the next sections.

15. Frozen shoulder

Frozen shoulder runs through three phases that merge into one another: a painful phase in which mobility gradually decreases; a stiff phase in which the pain subsides and the restriction dominates; and a thawing phase in which mobility returns over months.

On the natural course there is a Finnish study that followed people for between 2 and 27 years: 94 per cent regained their normal function without any procedure, and in the group that had received no treatment at all the condition lasted 15 months on average, with a range of 4 to 36 months [30]. This figure is the most important piece of information for those affected, and it is also hard to bear, because 15 months too is a long time.

On treatment, the studies give the following picture:

• In the early, painful phase the corticosteroid injection into the joint is the most effective single measure. A network meta-analysis found for it a clear short-term advantage in pain over no treatment or a sham treatment, and likewise over physiotherapy alone [28]. The combination of an early injection with a home programme of simple exercises and stretches performed best, with additional benefit in the medium term.
• A short course of oral corticosteroids (about 30 milligrams of prednisolone over three weeks, then tapered) can relieve severe pain [1].
• Exercise and manual therapy are sensible in the stiff phase, once the pain has subsided and the restriction of movement remains [1]. In the very painful early phase, severe pain can prevent participation in exercise, and the load is then reduced accordingly.
• A comparison of watchful waiting with physiotherapy in a randomised trial produced similar results for function after one year, with less effort for those who waited [31]. This result supports a calm approach and speaks against the idea that an intensive stretching programme can shorten the course.
• Procedures come into question only when a substantial restriction of movement persists beyond about 18 months [1]. The British UK FROST trial compared three paths at 35 hospitals: an early structured physiotherapy programme with a corticosteroid injection, manipulation under anaesthesia, and arthroscopic release of the capsule. After twelve months, arthroscopic release led on the questionnaire score (40.3 points), followed by manipulation under anaesthesia (38.3) and physiotherapy (37.2). The differences were statistically detectable and not clinically important; arthroscopic release carried the highest risk of complications, and manipulation under anaesthesia was the most economical option [29].

Three points can be drawn from this for everyday life. First, it helps to know the expected time frame, because expectation and patience follow from it. Second, it is worth discussing an injection in the painful early phase, because it can make sleep possible. Third, an exercise programme belongs with it, adapted to the phase, and its benefit lies in maintaining the remaining mobility and securing strength.

Two further points belong in the conversation: people with diabetes are affected considerably more often and often have a more stubborn course, and after one frozen shoulder there is an increased risk that the other shoulder will be affected too [1].

There is a separate, detailed article on frozen shoulder: Frozen shoulder – with the figures on its course, the question of how hard you may stretch, and the procedures compared.

16. A tear in the rotator cuff

The word «tear» creates the image of a sudden event. For the rotator cuff, this image applies to a minority.

Three situations are to be distinguished:

• The age-related tear without an injury. It develops over years and is found, as described above, in a large part of the pain-free population [5]. In most people such a tear is an incidental finding.
• The tear after an injury in younger people, often with immediate marked weakness. Here a surgical assessment is more likely to be recommended.
• The large tear with substantial loss of function, in which the arm can no longer be raised above shoulder height.

For the most common situation, the age-related tear without an injury, the evidence is well studied. The Cochrane review on surgery for tears summarises three trials with 258 participants in which surgery was compared with non-surgical treatment: after twelve months the difference in pain was 0.9 points on a scale of 0 to 10, at moderate certainty of evidence [26] – a difference below the threshold at which people experience a change as meaningful.

A Finnish randomised trial in people over 55 with small supraspinatus tears without an injury compared physiotherapy, physiotherapy with smoothing of the acromion, and physiotherapy with repair of the tendon. After more than five years the three groups did not differ meaningfully in pain and function [27]. A further investigation found no evidence that repair prevents the progression of osteoarthritis of the shoulder joint [1].

For physiotherapy, such a tear therefore rarely means the end of the possibilities. The remaining muscles of the cuff and the muscles guiding the shoulder blade can be trained, and many people achieve good function that way. A conversation about surgery is appropriate in young people, for tears after an injury, for large tears with marked loss of function, and when a carefully conducted exercise programme has shown no effect over several months.

17. Calcific tendinitis

Calcific tendinitis runs in episodes. An acute phase can be so painful that any movement of the arm becomes impossible; it usually subsides over days to a few weeks. Calcium deposits frequently dissolve on their own.

On treatment: in the acute phase, pain relief and a corticosteroid injection are to the fore. A Norwegian randomised trial with 220 participants compared ultrasound-guided lavage of the deposit with a corticosteroid injection, sham lavage with a corticosteroid injection, and sham lavage without corticosteroid. The corticosteroid injection was superior to the sham treatment at two and at six weeks, and the lavage brought no additional benefit [32]. For ultrasound and laser there are indications of a short-term benefit precisely in the presence of calcium deposits, at low certainty of evidence [19]. Shockwave therapy is not recommended according to the Cochrane review [20].

After the acute phase has subsided, the same approach applies as for subacromial pain: restore mobility, build strength, increase load.

18. Osteoarthritis of the shoulder joint and of the AC joint

For the shoulder joint the treatment at the outset is the same as for subacromial pain: explanation, adaptation of activities, simple painkillers, watchful waiting [1]. Studies comparing an exercise programme for osteoarthritis of the shoulder joint with a sham treatment or with watchful waiting are so far lacking. From experience with other joints and from the review it follows that a corticosteroid injection into the joint can provide short-term relief for two to six weeks. Hyaluronic acid has shown no advantage over saline in two randomised trials. Assessment by a surgeon is appropriate when the complaints persist or increase despite non-surgical treatment and function is substantially restricted; total joint replacement is preferred over partial replacement.

For the AC joint the complaints frequently recede on their own. Treatment consists of adapting activities, topical or oral anti-inflammatory drugs and, where appropriate, a corticosteroid injection. Surgery in which the outer end of the collarbone is removed remains reserved for persistent and precisely localised complaints that do not respond to non-surgical treatment [1].

19. Sleeping, working, sport

Sleep. Night pain is the greatest burden for many of those affected, and it can often be eased by simple means. What has proved useful: lying on the healthy side and placing the painful arm on a pillow in front of the body so that it does not fall forwards or backwards; when lying on the back, sliding a flat pillow under the upper arm so that the shoulder does not sink backwards; and, if the pain prevents falling asleep, an anti-inflammatory drug in the evening, after consulting your doctor. When night pain prevents sleep for weeks, that is a good reason to discuss a corticosteroid injection.

Work. People with physically demanding work and those who work a great deal above shoulder height are particularly affected [1]. A temporary adaptation makes sense: change the height of the work so that it takes place below shoulder height; spread heavy tasks over the day instead of completing them in one block; switch sides where that is possible. Complete sick leave is rarely the best route, because the arm then drops out of loading altogether and the return becomes more difficult as a result.

Sport. Sports with movements above the head – tennis, volleyball, handball, swimming, climbing – place particular load on the shoulder. Complete abstention is rarely necessary. A graded approach makes sense: first reduce frequency and volume, then leave out the most demanding movements, and only at the very end pause entirely. In strength training, painful exercises can usually be replaced without losing the training stimulus: bench press with a narrower grip or on a slight incline, pulling towards the body instead of pulling to the neck, replacing overhead pressing with pressing in the incline plane.

20. What else influences the course

A British study followed people referred to physiotherapy for shoulder pain and examined which characteristics are associated with a better outcome at six months [33]. The result is remarkable:

• Four characteristics were associated with a better course: less restriction at the start, the expectation of complete recovery through physiotherapy, greater confidence in one's own ability to deal with pain (technical term: pain self-efficacy), and less pain at rest.
• The findings of the physical examination that point to a particular structure were not associated with the course.

This observation explains why the guideline explicitly recommends addressing and working on fear, stress and low confidence [12]. It also explains why a good explanation at the outset is part of the treatment itself.

Two things follow for you as an affected person. First, the severity of your complaints at the start says little about the further course. Second, confidence and expectation can change, and they change most readily through experiencing that a movement which seemed impossible three weeks ago succeeds today. That is precisely a reason to record progress in writing.

A simple approach: choose two or three activities that matter to you and that are currently difficult – reaching something from the top shelf, putting on the seat belt, drying your hair. Rate these activities every two weeks on a scale of 0 to 10. Also note the resistance with which you perform your main exercise. These two records show progress more reliably than the pain on any single day.

21. Ten misunderstandings

• «An MRI will finally bring clarity.» In 96 out of 100 pain-free shoulders in the age group from 41 years upwards, at least one abnormality of the rotator cuff is found [5]. A finding alone therefore does not yet explain the pain.
• «A tear has to be stitched.» For the age-related tear without an injury, surgery and non-surgical treatment did not differ meaningfully after five years [27].
• «The operation makes more room and solves the problem.» Subacromial decompression was not superior to placebo surgery at twelve months, nor at ten years [23][24].
• «I have to spare the arm until it is better.» Prolonged sparing leads to loss of strength and to stiffness. The recommendation is to adapt demanding activities and to stay in movement [1].
• «If it hurts while exercising, I am damaging something.» Temporary pain during exercise does not stand in the way of success, as long as it subsides again within a day [17].
• «My posture is to blame.» A uniform malposition of the shoulder blade in people with shoulder pain could not be demonstrated [21].
• «The injection heals the shoulder.» The benefit of the corticosteroid injection is small and lasts a few weeks [22]. Its value lies in opening a window of time for the exercises.
• «With a frozen shoulder you have to stretch hard so that it opens up.» The course cannot be shortened by intensive stretching. In a randomised trial, watchful waiting achieved a similar level of function after one year as physiotherapy [31].
• «Ultrasound, laser and shockwave are modern methods and therefore effective.» For ultrasound and laser there are indications of a short-term benefit with calcium deposits, at low certainty [19]; for shockwave in subacromial pain the evidence speaks against it [20].
• «If it is still there after three months, it will never go away.» Around 80 per cent of those affected improve within a year [1], and in the ten-year observation close to 80 per cent had improved satisfactorily [7].

22. When you should get in touch

Seek medical assessment immediately:

• Fever together with a hot, swollen, very painful shoulder.
• Chest pain, breathlessness, nausea or sweating together with shoulder pain.
• Marked deformity or loss of strength after a fall or an injury.
• Increasing paralysis or numbness in the arm.

Arrange an appointment within the next few days:

• Unintended weight loss, night sweats or a history of cancer together with shoulder pain.
• Complaints in both shoulders at once, morning stiffness lasting over an hour, complaints in further joints.
• The arm can no longer be raised above shoulder height.

Discuss at the next opportunity:

• Night pain has been preventing sleep for several weeks.
• The complaints are worsening over weeks instead of staying the same or improving.
• A carefully conducted exercise programme has changed nothing after three months.
• Mobility is decreasing although you are exercising. This can point to a frozen shoulder that only shows itself over the course.

23. In summary

Shoulder pain is common, in the great majority of cases harmless, and improves within a year in about four out of five people affected. The most important task of the first examination is to exclude the rare serious causes and to clarify whether the pain comes from the shoulder at all. Once that has been done, the treatment is largely similar across the different forms.

An image at the outset as a rule brings no benefit, because abnormalities of the rotator cuff are almost always present in people from 40 years upwards, including without complaints. Imaging is appropriate with warning signs, after an injury, with increasing weakness and with complaints that are worsening.

Physiotherapy consists of an understandable explanation, an adaptation of everyday life and an exercise programme that includes external rotation and the muscles guiding the shoulder blade and that progresses over the weeks. The evidence that this improves strength is unambiguous; the evidence for an advantage in pain after one year is not, because most people have improved by then anyway. The largest trial on the subject shows that a single good advice session with a home programme works as well over twelve months as six supervised sessions.

Medication and corticosteroid injections have a limited but sensible role: they create a window of time in which movement and sleep become possible again. Subacromial decompression is not superior to placebo surgery in subacromial pain without a full-thickness tendon tear, not even after ten years. Frozen shoulder lasts on average about 15 months and recovers in the great majority of people without any procedure.

What you can influence most strongly yourself is the regularity of the exercises, the stepwise increase in load and the handling of your own expectations. These three points are precisely the ones most clearly associated with a good course in the studies.

References

All Digital Object Identifiers (DOIs) were checked individually against the Crossref register in August 2026. The links in the reference list lead via the DOI service to the publishers' pages. Some of these are located outside Switzerland and the EU. When you click on them, your IP address is transmitted to the provider concerned – on our own site this does not happen.

[1] Haas R, Ibounig T, Buchbinder R. Management of Shoulder Pain in Primary Care: A Review. JAMA Internal Medicine. 2026. Published online August 17, 2026. https://doi.org/10.1001/jamainternmed.2026.3135

[2] Lucas J, van Doorn P, Hegedus E, Lewis J, van der Windt D. A systematic review of the global prevalence and incidence of shoulder pain. BMC Musculoskeletal Disorders. 2022;23(1):1073. https://doi.org/10.1186/s12891-022-05973-8

[3] Lewis J, Mintken PE, McDevitt AW. What’s in a Name? The Case for Using “Rotator Cuff–Related Shoulder Pain” in Clinical Practice. Journal of Orthopaedic & Sports Physical Therapy. 2025;55(7):451–453. https://doi.org/10.2519/jospt.2025.13405

[4] Zadro JR, Michaleff ZA, O'Keeffe M, et al. How do people perceive different labels for rotator cuff disease? A content analysis of data collected in a randomised controlled experiment. BMJ Open. 2021;11(12):e052092. https://doi.org/10.1136/bmjopen-2021-052092

[5] Ibounig T, Järvinen TLN, Raatikainen S, et al. Incidental Rotator Cuff Abnormalities on Magnetic Resonance Imaging. JAMA Internal Medicine. 2026;186(4):406–414. https://doi.org/10.1001/jamainternmed.2025.7903

[6] Sanders S, Ibounig T, Haas R, et al. Rotator Cuff Imaging Abnormalities in Asymptomatic Shoulders: A Systematic Review. Journal of Orthopaedic & Sports Physical Therapy. 2025;55(12):736–751. https://doi.org/10.2519/jospt.2025.13611

[7] Dorrestijn O, Greving K, van der Veen WJ, van der Meer K, Diercks RL, Winters JC, Stevens M. Patients with shoulder complaints in general practice: consumption of medical care. Rheumatology. 2011;50(2):389–395. https://doi.org/10.1093/rheumatology/keq333

[8] Reilingh ML, Kuijpers T, Tanja-Harfterkamp AM, van der Windt DA. Course and prognosis of shoulder symptoms in general practice. Rheumatology. 2008;47(5):724–730. https://doi.org/10.1093/rheumatology/ken044

[9] Zhang Z, Ferreira GE, Downes JS, et al. The effectiveness of education for people with shoulder pain: A systematic review. Musculoskeletal Science and Practice. 2025;75:103246. https://doi.org/10.1016/j.msksp.2024.103246

[10] Hopewell S, Keene DJ, Marian IR, et al. Progressive exercise compared with best practice advice, with or without corticosteroid injection, for the treatment of patients with rotator cuff disorders (GRASP): a multicentre, pragmatic, 2 × 2 factorial, randomised controlled trial. The Lancet. 2021;398(10298):416–428. https://doi.org/10.1016/S0140-6736(21)00846-1

[11] Steuri R, Sattelmayer M, Elsig S, Kolly C, Tal A, Taeymans J, Hilfiker R. Effectiveness of conservative interventions including exercise, manual therapy and medical management in adults with shoulder impingement: a systematic review and meta-analysis of RCTs. British Journal of Sports Medicine. 2017;51(18):1340–1347. https://doi.org/10.1136/bjsports-2016-096515

[12] Desmeules F, Roy JS, Lafrance S, et al. Rotator Cuff Tendinopathy Diagnosis, Nonsurgical Medical Care, and Rehabilitation: A Clinical Practice Guideline. Journal of Orthopaedic & Sports Physical Therapy. 2025;55(4):235–274. https://doi.org/10.2519/jospt.2025.13182

[13] Zhang B, Raguzzi IA, Dupuis F, Gianola S, Morgan-Daniel J, Roy JS, Pozzi F. Addressing Shoulder Weakness in Individuals With Rotator Cuff–Related Shoulder Pain: A Systematic Review With Meta-analysis. Journal of Orthopaedic & Sports Physical Therapy. 2026;56(2):67–84. https://doi.org/10.2519/jospt.2025.13445

[14] Lafrance S, Charron M, Dubé MO, et al. The Efficacy of Exercise Therapy for Rotator Cuff–Related Shoulder Pain According to the FITT Principle: A Systematic Review With Meta-analyses. Journal of Orthopaedic & Sports Physical Therapy. 2024;54(8):499–512. https://doi.org/10.2519/jospt.2024.12453

[15] Naunton J, Street G, Littlewood C, Haines T, Malliaras P. Effectiveness of progressive and resisted and non-progressive or non-resisted exercise in rotator cuff related shoulder pain: a systematic review and meta-analysis of randomized controlled trials. Clinical Rehabilitation. 2020;34(9):1198–1216. https://doi.org/10.1177/0269215520934147

[16] Holmgren T, Bjornsson Hallgren H, Oberg B, Adolfsson L, Johansson K. Effect of specific exercise strategy on need for surgery in patients with subacromial impingement syndrome: randomised controlled study. BMJ. 2012;344:e787. https://doi.org/10.1136/bmj.e787

[17] Smith BE, Hendrick P, Smith TO, et al. Should exercises be painful in the management of chronic musculoskeletal pain? A systematic review and meta-analysis. British Journal of Sports Medicine. 2017;51(23):1679–1687. https://doi.org/10.1136/bjsports-2016-097383

[18] Page MJ, Green S, McBain B, et al. Manual therapy and exercise for rotator cuff disease. Cochrane Database of Systematic Reviews. 2016;(6):CD012224. https://doi.org/10.1002/14651858.CD012224

[19] Page MJ, Green S, Mrocki MA, et al. Electrotherapy modalities for rotator cuff disease. Cochrane Database of Systematic Reviews. 2016;(6):CD012225. https://doi.org/10.1002/14651858.CD012225

[20] Surace SJ, Deitch J, Johnston RV, Buchbinder R. Shock wave therapy for rotator cuff disease with or without calcification. Cochrane Database of Systematic Reviews. 2020;(3):CD008962. https://doi.org/10.1002/14651858.CD008962.pub2

[21] Ratcliffe E, Pickering S, McLean S, Lewis J. Is there a relationship between subacromial impingement syndrome and scapular orientation? A systematic review. British Journal of Sports Medicine. 2014;48(16):1251–1256. https://doi.org/10.1136/bjsports-2013-092389

[22] Mohamadi A, Chan JJ, Claessen FMAP, Ring D, Chen NC. Corticosteroid Injections Give Small and Transient Pain Relief in Rotator Cuff Tendinosis: A Meta-analysis. Clinical Orthopaedics & Related Research. 2017;475(1):232–243. https://doi.org/10.1007/s11999-016-5002-1

[23] Karjalainen TV, Jain NB, Page CM, et al. Subacromial decompression surgery for rotator cuff disease. Cochrane Database of Systematic Reviews. 2019;(1):CD005619. https://doi.org/10.1002/14651858.CD005619.pub3

[24] Kanto K, Bäck M, Ibounig T, et al. Arthroscopic subacromial decompression versus placebo surgery for subacromial pain syndrome: 10 year follow-up of the FIMPACT randomised, placebo surgery controlled trial. BMJ. 2025;391:e086201. https://doi.org/10.1136/bmj-2025-086201

[25] Vandvik PO, Lähdeoja T, Ardern C, et al. Subacromial decompression surgery for adults with shoulder pain: a clinical practice guideline. BMJ. 2019;364:l294. https://doi.org/10.1136/bmj.l294

[26] Karjalainen TV, Jain NB, Heikkinen J, Johnston RV, Page CM, Buchbinder R. Surgery for rotator cuff tears. Cochrane Database of Systematic Reviews. 2019;(12):CD013502. https://doi.org/10.1002/14651858.CD013502

[27] Kukkonen J, Ryösä A, Joukainen A, Lehtinen J, Kauko T, Mattila K, Äärimaa V. Operative versus conservative treatment of small, nontraumatic supraspinatus tears in patients older than 55 years: over 5-year follow-up of a randomized controlled trial. Journal of Shoulder and Elbow Surgery. 2021;30(11):2455–2464. https://doi.org/10.1016/j.jse.2021.03.133

[28] Challoumas D, Biddle M, McLean M, Millar NL. Comparison of Treatments for Frozen Shoulder: A Systematic Review and Meta-analysis. JAMA Network Open. 2020;3(12):e2029581. https://doi.org/10.1001/jamanetworkopen.2020.29581

[29] Rangan A, Brealey SD, Keding A, et al. Management of adults with primary frozen shoulder in secondary care (UK FROST): a multicentre, pragmatic, three-arm, superiority randomised clinical trial. The Lancet. 2020;396(10256):977–989. https://doi.org/10.1016/S0140-6736(20)31965-6

[30] Vastamäki H, Kettunen J, Vastamäki M. The Natural History of Idiopathic Frozen Shoulder: A 2- to 27-year Followup Study. Clinical Orthopaedics & Related Research. 2012;470(4):1133–1143. https://doi.org/10.1007/s11999-011-2176-4

[31] Martin SD, Dean MC, Eberlin CT, et al. AAOS 2024 best paper in the shoulder and elbow classification: watchful waiting provides higher value with similar functional outcomes to physical therapy for frozen shoulder: a prospective randomized controlled trial. Journal of Shoulder and Elbow Surgery. 2025;34(6):e455-e467. https://doi.org/10.1016/j.jse.2024.09.017

[32] Moosmayer S, Ekeberg OM, Hallgren HB, et al. Ultrasound guided lavage with corticosteroid injection versus sham lavage with and without corticosteroid injection for calcific tendinopathy of shoulder: randomised double blinded multi-arm study. BMJ. 2023;383:e076447. https://doi.org/10.1136/bmj-2023-076447

[33] Chester R, Jerosch-Herold C, Lewis J, Shepstone L. Psychological factors are associated with the outcome of physiotherapy for people with shoulder pain: a multicentre longitudinal cohort study. British Journal of Sports Medicine. 2018;52(4):269–275. https://doi.org/10.1136/bjsports-2016-096084