Get the phase right and you cut the worst months down. A cortisone shot into the joint during the painful phase usually returns nights to baseline within a few weeks; gentle pain-respecting movement through the stiff phase keeps you on track; full range comes back during thawing. The three ways readers make it worse: skipping the shot, forcing range of motion through pain, or chasing surgery in the first year.
The shoulder joint sits inside a soft tissue bag called the capsule. In a frozen shoulder, that capsule gets inflamed and starts laying down scar tissue inside itself β the same kind of tissue that forms a Dupuytren's contracture in the palm Bunker and Anthony 1995. The scar tightens and shortens the bag from the inside, and the inflammation drives the pain. External rotation β turning the hand outward, like reaching into the back seat β usually goes first, then lifting the arm overhead, then reaching behind your back.
It's not a torn muscle, it's not arthritis, and it's not your rotator cuff. The lesion is the capsule itself, which is why pulling harder on a frozen shoulder doesn't unstick it the way you'd hope. The capsule moves through inflammation first, then mature scar, then slow remodelling β three biological windows, three different things that help.
What actually works
The most-cited mistake is treating a frozen shoulder like a stiff joint that needs to be stretched out. The opposite is closer to true: in the painful phase, the less you force, the faster you recover.
The companion finding from a placebo-controlled trial: a single cortisone injection into the joint combined with physiotherapy beats physiotherapy alone for pain and function at six weeks Carette et al. 2003. A 2017 meta-analysis pooling multiple trials confirms the short-term injection edge Sun et al. 2017. The gap narrows by six months β the natural course catches up β but the painful months in between are real, and the injection takes the edge off them.
For shoulders that are still stuck a year in, the biggest trial yet β UK FROST, 503 patients across 35 hospitals β randomised people to three arms: surgery (arthroscopic capsular release), manipulation under anaesthesia, and ongoing physiotherapy-plus-injection. At twelve months, all three produced roughly equivalent results, and the surgery arm carried more complications Rangan et al. 2020. The takeaway: the natural course is doing most of the work, the injection takes the edge off the worst months, and surgery is rarely worth its complication rate in the first year.
What to do, by phase
Pick the description that fits where you are right now. The boundaries are fuzzy and the phases overlap, so use it as a guide, not a strict rule.
Painful (freezing) phase β roughly months 0 to 9
Pain is the dominant symptom. The giveaway is night pain: you can't lie on the shoulder, and you can't quite lie off it either. Stiffness is building but isn't the worst of it yet.
The job is pain control and protecting the joint. The most useful single thing you can do is get a cortisone shot into the joint from a clinician who uses ultrasound guidance β the meta-analyses are clear that it helps faster than physiotherapy alone for short-term pain Sun et al. 2017. Keep the shoulder moving gently within the pain-free range; don't force it.
Frozen (stiff) phase β roughly months 9 to 15
Pain settles to the background; stiffness becomes the main problem. You can't reach into the back seat, you can't fasten a bra strap, you can't put a coat on without a struggle.
This is when physiotherapy earns its keep. A therapist can step up the intensity: stretches held for 20 to 30 seconds, hands-on joint mobilisation, and active range work. The 2013 American physical therapy guidelines call this "irritability-graded" progression β the calmer the capsule, the more you can push Kelley et al. 2013. A second injection can help if pain returns.
Thawing phase β roughly months 15 to 24+
Pain is gone or background; motion is steadily returning. One morning you reach the top shelf without thinking about it.
Strengthening goes in alongside the stretching now β the rotator cuff and the muscles around the shoulder blade have weakened from months of disuse and need to come back online. Full motion typically returns by the two-year mark, though a long-term cohort study found around four in ten people with some mild residual stiffness at four years Hand et al. 2008.
When to be careful
Diabetes plus a cortisone shot is the most common landmine. The injection raises blood sugar for two to five days afterward β typically by 50 to 200 mg/dL β so if you're insulin-dependent or running tight glycemic control, plan for it. Don't skip the shot if your shoulder is genuinely frozen; just coordinate with whoever manages your diabetes Zreik et al. 2016.
The other landmine is the opposite of the first one: in the painful phase, end-range stretching and aggressive hands-on joint cranking make things worse Diercks and Stevens 2004. If a physical therapist is pushing your shoulder past the pain threshold during the early months, find a different therapist.
What most people get wrong
"Work through the pain." The most damaging advice. The painful phase is an inflammatory phase; cranking the capsule prolongs the inflammation and extends the disability. In the Dutch trial above, the group that pushed past pain ended up worse off at two years than the group that didn't Diercks and Stevens 2004.
"It always fully resolves on its own." Mostly true, with a footnote. A long-term cohort found roughly four in ten people had some mild residual stiffness at four years Hand et al. 2008. Letting it ride untreated also means twelve to eighteen months of unnecessarily bad sleep.
"I need surgery." Rarely in the first year. The UK FROST trial β 503 patients, three arms β found that aggressive surgery was clinically equivalent to manipulation under anaesthesia and to staying with physiotherapy plus injection at twelve months, and carried more complications Rangan et al. 2020.
"It's my rotator cuff." The clue is passive motion. In a rotator cuff problem, someone else can lift your arm even when you can't. In frozen shoulder, nobody can lift your arm past the restricted range β the capsule itself is the brake, and external force can't override it.
Who gets it
The typical patient is a woman between 40 and 60. Men get it too, just less often. The single strongest risk factor is diabetes: diabetics have roughly five times the prevalence of non-diabetics, and around three in ten people who show up with a frozen shoulder turn out to have diabetes, sometimes diagnosed because of the shoulder Zreik et al. 2016. A 2021 genetic study using Mendelian randomisation also pointed at diabetes as a causal contributor, not just a co-traveller Green et al. 2021.
Other elevated-risk groups: hypothyroidism, prior shoulder trauma or surgery, prolonged immobilisation (a cast, a sling), Dupuytren's contracture in the hand. Diabetic frozen shoulder also tends to be the worst kind β longer course, more likely to involve both shoulders eventually, less responsive to every conservative treatment. If you have diabetes and your shoulder is freezing up, treat it earlier rather than later.
If you ignore it
You don't die from a frozen shoulder; you lose roughly a year of normal sleep and another six months of a normal arm. The first time you try to grab the seatbelt with the bad side and end up swearing in your driveway is week six. By month three your partner has stopped asking if you want to lift things together because the answer is always no. By month six you're lying in bed at three in the morning trying to find a position that doesn't feel like someone's twisting a knife in your shoulder; you're not finding it. The PT person, the meeting, the lunch β you start cancelling them, then resenting yourself for cancelling.
People at work start adjusting around you. Someone takes the heavy side of the box; someone reaches across you for the file. You're aware that you've become the person who needs adjusting around, and it sits in the same place that the pain does. Untreated, that goes on for fifteen months on average, sometimes thirty VastamΓ€ki et al. 2012. Most people get their motion back eventually; about four in ten end up with some residual stiffness years later, mostly mild but real Hand et al. 2008. The cost of doing nothing isn't the residual stiffness β it's the year of three a.m. wakeups you didn't need to have.
If you respond well
The injection lands and within two or three weeks the three-a.m. wakeups stop Carette et al. 2003. That alone changes the next month: you start dreaming again, your partner stops asking how you slept because the answer isn't "badly" anymore. Six weeks in, you're not pain-free but the dread has lifted; you can sleep on either side, you can carry a grocery bag, you can put a jacket on. By six months, with phase-appropriate physiotherapy underneath, you can reach the top shelf and you've stopped flinching when someone reaches across you.
At one year you're 80 to 90 percent of the way back; at two years you're somewhere you'd call normal. The grind didn't disappear β the natural course is doing its work in the background no matter what β but the worst months got compressed, and you got your nights back early Rangan et al. 2020. The mood lift from sleep coming back, and from no longer being the person who can't carry things, is the part nobody warns you about until you've been through it.
Adjacent
The other big shoulder-pain cause is rotator-cuff disease β different mechanism, different exam, different treatment. Diabetes management sits upstream of the risk itself; if you don't know your numbers, find them out. Sleep posture and pillow setup matter a lot during the painful months β sleeping on the unaffected side with a pillow propping the painful arm is the standard move. Long-term, scapular strength and posture work reduce the chance of secondary stiffness if you've had a frozen shoulder once and want to protect the other side.
- β Frozen shoulder is the stiffness-dominant cousin of rotator cuff pain β telling them apart changes the whole treatment.
- β Diabetes sharply raises the odds of a frozen shoulder, and the course can run longer β worth knowing if sugars are high.
- β Frozen shoulder travels with carpal tunnel - both are soft-tissue conditions far more common in diabetes and low thyroid.
- β An underactive thyroid raises your odds of a frozen shoulder; if it freezes for no clear reason, it's worth checking your TSH.
Substance and claimed effects
Frozen shoulder (adhesive capsulitis) is a self-limiting fibroproliferative contracture of the glenohumeral joint capsule, marked by progressive pain and global loss of both active and passive shoulder motion, with external rotation typically the first and most severely restricted plane Le et al. 2017Ramirez 2019. Lifetime prevalence is approximately 2-5% in the general population, with peak onset between ages 40 and 60, female predominance roughly 1.4:1, and contralateral involvement in 6-17% of cases Le et al. 2017Hsu et al. 2011. The classical clinical course is three overlapping phases β freezing (painful, 2-9 months), frozen (stiffness-dominant, 4-12 months), and thawing (gradual recovery, 5-24 months) β totalling 1-3 years untreated Reeves 1975Hsu et al. 2011. Effects this entry scores across: large short-term quality-of-life and sleep impacts from disabling night pain and arm-use loss, secondary mood and energy effects, and the phase-specific benefit of intra-articular corticosteroid for pain and physical therapy for motion. No effect on appearance or longevity.
Evidence by addressing question
Mechanism
The driving lesion is fibroblastic and myofibroblastic proliferation in the glenohumeral capsule, particularly the rotator interval (coracohumeral and superior glenohumeral ligaments), with histology comparable to Dupuytren's contracture Bunker and Anthony 1995. Type III collagen deposition, mast-cell infiltration, neoangiogenesis, neoneuralisation, and elevated cytokine signalling (TGF-Ξ², IL-6, MMPs) progressively thicken and shorten the capsule, obliterating the axillary recess Le et al. 2017Lewis 2015. Inflammation dominates the freezing phase and explains why intra-articular corticosteroid is most useful early; mature fibrosis dominates the frozen phase and explains why mechanical interventions (manipulation, capsular release, hydrodilatation) are reserved for the stiff-dominant period Lewis 2015Cho et al. 2019. The 2021 GWAS by Green et al. identified five risk loci including a variant near WNT7B (OR 1.20, p = 5Γ10-29) and provided Mendelian-randomisation evidence that type 1 diabetes is a causal risk factor (OR 1.03, p = 3Γ10-6), suggesting a glycemic rather than mechanical pathway Green et al. 2021.
Evidence (does treatment work?)
Three large evidence streams: (1) Intra-articular corticosteroid injection produces clinically meaningful short- to mid-term pain reduction and faster recovery of motion versus placebo or physiotherapy alone. Carette et al.'s 4-arm placebo-controlled RCT (n=93) showed steroid injection plus physiotherapy was significantly better than physiotherapy alone, placebo injection plus physiotherapy, or placebo alone at 6 weeks, with the steroid-arm differences attenuating by 6 months Carette et al. 2003. Sun et al.'s 2017 meta-analysis confirms steroid injections beat physiotherapy at short-term endpoints (4-6 weeks) on pain and function, with convergence at 6 months Sun et al. 2017. (2) For physiotherapy intensity, the landmark Diercks & Stevens prospective study (n=77, 24-month follow-up) found 89% of the "supervised neglect" group (pain-free range, pendulum exercises, no forced stretching) achieved Constant scores β₯80 versus 63% of the intensive-stretching group, with the supervised-neglect group recovering faster Diercks and Stevens 2004. The 2014 Cochrane review of manual therapy and exercise found low-to-moderate quality evidence that combining mobilisation/exercise with steroid injection beats either alone Page et al. 2014. The 2013 APTA clinical practice guideline grades stretching tailored to irritability level as the standard, with intensity progressing only as the capsule transitions from inflamed to fibrotic Kelley et al. 2013. (3) For refractory disease, the UK FROST three-arm RCT (n=503, 35 sites) compared early structured physiotherapy with steroid injection, manipulation under anaesthesia (MUA) with steroid, and arthroscopic capsular release (ACR) with MUA; Oxford Shoulder Scores at 12 months were 37.2, 38.3, and 40.3 respectively β clinically equivalent, with ACR carrying higher complication rates and MUA the most cost-effective option Rangan et al. 2020. Hydrodilatation (saline distension of the capsule with steroid) shows short-term benefit comparable to or modestly better than steroid alone in meta-analyses, but the comparative trial base is small and the procedure was excluded from UK FROST Saltychev et al. 2018. The natural-history baseline: VastamΓ€ki et al.'s 9-year mean follow-up of 103 patients (untreated and conservative-treatment arms) found 94% recovered range of motion matching the contralateral shoulder, with untreated duration averaging 15 months (range 4-36) VastamΓ€ki et al. 2012. Hand et al.'s longer-term follow-up reports residual symptoms in 40% at 4+ years, though most are mild Hand et al. 2008.
Protocol
Phase-stratified care is the consensus: in the freezing (high-irritability) phase, prioritise pain control β intra-articular corticosteroid injection (most studied: 40 mg triamcinolone), NSAIDs as adjunct, and pain-free home exercises (pendulum, scapular setting, motion within the painless range); avoid forced end-range stretching Carette et al. 2003Kelley et al. 2013Diercks and Stevens 2004. In the frozen (moderate-irritability) phase, transition to progressive mobilisation β passive and active-assisted stretching at sustained holds, joint mobilisation, capsular stretches; steroid injection may still help if pain persists Kelley et al. 2013Cho et al. 2019. In the thawing (low-irritability) phase, prioritise strengthening and end-range work to restore full function. Surgical escalation (MUA, ACR, hydrodilatation) is considered after 6-12 months of failed conservative management or refractory disability, with patient preference and risk profile guiding choice β UK FROST establishes that none of the three care pathways is clearly superior at 12 months Rangan et al. 2020.
Contraindications
Corticosteroid injection is contraindicated in active joint infection and relatively contraindicated in poorly controlled diabetes given transient post-injection hyperglycemia (typically 50-200 mg/dL elevation lasting 2-5 days) β counsel diabetic patients to monitor and adjust glycemic control around the injection Ramirez 2019Zreik et al. 2016. Aggressive end-range stretching in the freezing phase is associated with worse outcomes in the Diercks/Stevens prospective comparison and is contraindicated by the 2013 APTA CPG when irritability is high Diercks and Stevens 2004Kelley et al. 2013. MUA carries small but real risks of humeral fracture, rotator cuff tear, dislocation, and brachial plexus injury Yip et al. 2018. Secondary capsulitis from prior shoulder surgery, fracture, or rotator-cuff pathology requires different management than primary idiopathic disease.
Misconceptions
(1) "Working through the pain" speeds recovery β contradicted by Diercks & Stevens, where the supervised-neglect (pain-respecting) arm outperformed intensive stretching 89% vs 63% on Constant scores at 24 months Diercks and Stevens 2004. (2) "It always fully resolves on its own" β Hand et al. report symptoms persisting in around 40% at 4+ years, mostly mild but not always trivial; UK FROST also shows substantial 12-month residual disability without treatment Hand et al. 2008Rangan et al. 2020. (3) "Surgery is the gold standard for stuck shoulders" β UK FROST showed clinical equivalence among physiotherapy+injection, MUA, and ACR at 12 months on patient-reported outcomes Rangan et al. 2020. (4) "It's a rotator cuff problem" β the lesion is capsular, not tendinous; MRI distinguishes them and clinical exam shows global passive ROM loss in frozen shoulder versus selective active-only weakness in rotator-cuff disease Ramirez 2019.
Audience
Women aged 40-60 carry the highest population risk; diabetics of either type have roughly 5x the prevalence of non-diabetics, with a 13.4% point prevalence in diabetic populations versus around 2-5% in the general population β and 30% of frozen-shoulder presentations carry a diabetes diagnosis Zreik et al. 2016. Other elevated-risk groups: thyroid dysfunction (especially hypothyroidism, ~27% versus 11% in controls), Dupuytren's contracture, prior shoulder trauma or surgery, prolonged immobilisation, and possibly cardiovascular disease and Parkinson's Wang et al. 2013Ramirez 2019. Diabetic frozen shoulder is more refractory: bilateral involvement is more common, duration is longer, and response to all conservative measures is more variable Zreik et al. 2016.
Failure modes
(1) Pushing range-of-motion exercises beyond pain in the freezing phase, prolonging inflammation. (2) Treating it as rotator-cuff impingement and getting a subacromial injection β meta-analysis shows intra-articular placement is significantly better for pain in frozen shoulder than subacromial Ramirez 2019. (3) Stopping physiotherapy after the injection feels good β pain relief without continued mobilisation lets stiffness mature. (4) Expecting a single injection to be curative β most protocols allow 2-3 injections at 6-8 week intervals if pain returns Cho et al. 2019. (5) Going to MUA or capsular release too early β UK FROST shows no surgical advantage at 12 months versus structured physiotherapy plus injection Rangan et al. 2020.
Stakes
Untreated, the median duration is 15-30 months with severe night pain and significant disability during the freezing and frozen phases VastamΓ€ki et al. 2012Reeves 1975. Sleep disturbance is reported in up to 81.5% of shoulder-pain populations Le et al. 2017. Chronic pain plus sleep loss plus loss of arm function drives a pain-sleep-mood triangle that increases depression and anxiety risk in this population, even though frozen shoulder itself is not life-threatening. Persistent functional limitation at 4+ years occurs in roughly 40% of those who go without treatment, though most residuals are mild Hand et al. 2008.
Payoff
Phase-appropriate management β early steroid injection in the painful phase, graded mobilisation through the frozen phase, full-range work in the thawing phase β measurably shortens disability and reduces pain severity in the short and mid term, even if the 12-month equipoise of UK FROST holds for surgical and conservative arms Carette et al. 2003Sun et al. 2017Rangan et al. 2020. Night pain returning to baseline within 2-4 weeks of injection is a commonly observed pattern in steroid-injection arms Carette et al. 2003. Restored sleep, restored arm-use in dressing/grooming/work, and avoidance of the most disabling months of the natural history are the practical wins.
Out-of-scope
Related but separate: rotator cuff tendinopathy and tears, calcific tendinopathy, shoulder impingement syndrome, glenohumeral osteoarthritis, cervical radiculopathy referring to the shoulder, post-stroke shoulder. Diabetes management and thyroid screening are upstream of risk modification. Sleep hygiene during the painful phase and ergonomic workplace adaptation are pragmatic adjuncts.
Credibility range
Optimist case
The natural history and the phase-stratified treatment ladder are among the better-described entities in musculoskeletal medicine. Multiple RCTs converge on intra-articular corticosteroid working in the painful phase; multiple long-term cohorts converge on 90%+ recovery of motion with conservative care; UK FROST is a high-quality multicentre trial that establishes equipoise across plausible pathways. A patient who recognises the phase model, gets an intra-articular steroid injection within the first few months, and follows a phase-appropriate exercise program will measurably reduce their worst months of disability and almost certainly return to normal or near-normal function.
Skeptic case
The condition is self-limiting; most of the apparent treatment effect at long horizons is regression toward the mean. UK FROST's clinical equivalence at 12 months across three very different interventions is the cleanest expression of this skepticism β if surgery, manipulation, and physiotherapy-plus-injection all converge, the active ingredient may largely be time. The Cochrane manual-therapy review found low-quality evidence overall Page et al. 2014. Diabetic frozen shoulder responds less well to everything and Hand et al.'s 40% residual rate undermines the "always resolves" framing Hand et al. 2008. The strongest skeptic point: aggressive treatment of an inflamed capsule is harmful (Diercks/Stevens) and many recommended modalities (ultrasound, TENS, laser, capsular distension) have weak comparative evidence.
Author's call
The condition is real, the phase model is real, and steroid injection in the freezing phase has solid short-term RCT support; physiotherapy is mostly about avoiding the wrong things (aggressive stretching of an inflamed capsule) and progressively restoring motion as inflammation settles. Surgical escalation is reasonable only after 6-12 months of failed conservative care and even then UK FROST shows roughly equivalent 12-month outcomes. The article should lean confidently into the phase model and the injection-in-the-painful-phase recommendation, while being honest that the natural history is the dominant force and that aggressive over-treatment can extend disability. Evidence score 4 β strong RCT base for the core moves, but long-term equivalence and residual-disability data temper a 5.
Stakeholders and incentives
- Orthopaedic surgeons β incentive to operate (MUA, ACR); UK FROST is the strongest counter-incentive evidence published in their own literature, partly de-escalating surgical volume.
- Physiotherapists β generally aligned with the phase-stratified, irritability-graded model (APTA CPG); some legacy practitioners still over-stretch the painful phase.
- Interventional radiologists / rheumatologists β incentive to inject; hydrodilatation in particular is more profitable than simple steroid injection but evidence advantage is small.
- Endocrinologists / diabetes care β relevant given the 5x risk ratio; tighter glycemic control is hypothesised but not proven to reduce frozen-shoulder risk or improve recovery.
- Patients β high incentive to "do something" given severity of night pain; vulnerable to under-evidenced modalities (ultrasound, laser, manipulation chiropractic).
Population variability
Female 40-60 carries the highest baseline risk. Diabetic populations have ~5x prevalence, longer duration, and worse response to conservative therapy; type 1 and type 2 are roughly equivalent in risk, with type 1 over-represented in refractory cases Zreik et al. 2016Green et al. 2021. Hypothyroid and Parkinson's populations have elevated risk. Post-surgical and post-trauma secondary capsulitis behaves somewhat differently than idiopathic β often longer course and less responsive to injection. Contralateral involvement at some point occurs in 6-17% of patients; recurrence in the same shoulder is rare Hsu et al. 2011.
Knowledge gaps
The optimal injection site (intra-articular vs subacromial vs rotator interval), the value of repeated injections versus single, the comparative advantage of hydrodilatation over plain intra-articular steroid, and the long-term equivalence question UK FROST opened all remain partly unresolved. Whether tight glycemic control in diabetic patients alters frozen-shoulder risk or recovery has not been tested in a trial. The mechanism by which type 1 diabetes causally raises risk β glycation of capsular collagen versus shared genetic susceptibility β is hypothesised but not proven. No disease-modifying pharmacotherapy exists; collagenase trials are early.
Scope is the substance β adhesive capsulitis β and its meaningful consequences across pain, range of motion, sleep, mood, and the phase-specific roles of intra-articular injection and physiotherapy, matching the brief. Treated as a condition entry under respond: the reader's win comes from recognising their phase and selecting the right intervention for it.
- Hydrodilatation deliberately not surfaced in the article. The evidence base is modest (Saltychev 2018 meta) and UK FROST explicitly excluded it. Adding it would lengthen protocol without changing what most readers should do; left in the research dossier as background. Future entry candidate if the comparative literature firms up.
- MUA vs ACR collapsed into "surgery." The within-surgery distinction is real but secondary to the more important reader takeaway (UK FROST equipoise). A separate decision-support entry on surgical escalation for shoulder stiffness could carry that detail.
- Focus dimension dropped to 0. The night-pain β sleep-debt β cognitive-fog chain is real but barely scorable; honest about a non-effect rather than scoring 1 for a downstream-of-sleep ripple. Energy kept at 2 because the felt-experience payoff section earns it.
- Audience restricted to 40-59 and 60+. Frozen shoulder is rare before 40 in idiopathic form; the rare younger diabetic case is acknowledged in the audience section but the surfacing decision favours the typical-reader case.
- Cost burden 2 is the global-typical-reader case (a few PT sessions plus 1-2 injections). Cost in single-payer systems is lower; private-insurance/uninsured US cost is higher. Could not be split cleanly without making the score region-specific.
- Diabetes interaction is treated as a clinical caveat inside contraindications rather than a formal
diabetes-medicationcontraindication token, because the entry doesn't tell the reader to inject themselves β it tells them to seek a clinician, who will manage the glycemic side. Reconsider if the meta tokens are intended to surface clinician-managed interactions too. - Future link candidates (when these exist): rotator cuff tendinopathy, diabetes glycemic control, sleep posture for shoulder/back pain, scapular strength and posture work.
- Diercks and Stevens 2004 carries a lot of weight in the article. It's the cleanest contrast for the "don't push through pain" claim. Replicated in spirit by APTA 2013 CPG. If a more recent RCT with larger n appears, swap or supplement.
Adhesive Capsulitis (Frozen Shoulder)
Get the right care in the right phase and you cut the worst months of pain and disability roughly in half. Most of the wellness win is here.
A painful frozen shoulder hammers sleep β you can't lie on it, you can't lie off it. A cortisone shot in the painful phase usually gives nights back within weeks.
Physical therapy and possibly an injection or two β typically a few hundred to about a thousand dollars over the course, less if insurance covers it.
Daily home exercises for months, plus appointments. A real commitment, not a lifestyle overhaul.
Several solid trials. The phase model and the role of injection in the painful phase are well-established; long-term comparisons between treatment paths are murkier.
When night pain stops wrecking sleep and you stop bracing the arm all day, daytime energy comes back. Real, but downstream.
Months of disabling pain plus an arm you can't use grinds mood down. Solving the shoulder lifts a load you didn't realise was there.