What Works
Tier 1 — Strong Evidence HIGH
Exercise Prescription
Tier 2 + 3 — Moderate / Emerging
Tier 2 — Moderate Evidence MODERATE
Tier 3 — Emerging EMERGING
What Doesn't Work
- Corticosteroid injection. No role in primary instability — no inflammatory pathology, and chondrotoxicity concerns.
- Indefinite bracing or activity avoidance for young contact athletes. Delays inevitable surgical decision while damaging bony architecture progressively (Rutgers 2022).
- "Strengthening exercises" without phase progression criteria. Vague rehab without explicit criteria-based gates underperforms structured protocols.
- Single-test diagnosis. No single physical exam test reaches sufficient diagnostic accuracy alone. Cluster + targeted imaging is mandatory.
- Time-based RTP at a fixed month. Kelley 2022 cohort: 22% recurrence with time-only vs 5% with criteria-based testing.
Return to Training
- Pain-free full active and passive range of motion in all planes
- Strength symmetry ≥90% Limb Symmetry Index by handheld or isokinetic dynamometry
- Negative apprehension test at 90° abduction–external rotation
- Sport-specific or job-specific drill completion without apprehension or pain
- WOSI score within MCID of pre-injury baseline (or above clinically defined return threshold)
- Tampa Scale of Kinesiophobia score below threshold (typically <37); SI-RSI within validated range if applicable
- No recurrence of dislocation or subluxation during the rehab progression
- Patient subjectively confident in the position and pattern that caused the original injury
Red Flags — Refer Urgently
If any of these are present, do not start a rehab program — get assessed in person, fast.
- Locked dislocation that will not reduce after two attempts. Soft-tissue interposition or neurovascular entrapment. A&E for reduction with imaging.
- Pulselessness, distal pallor, or a cold limb. Axillary artery injury, rare but reported especially in the elderly. A&E vascular surgery immediately.
- Persistent dense numbness or weakness in the deltoid distribution beyond the immediate post-reduction period. Axillary nerve neurapraxia or brachial plexus traction injury. Urgent neurology / orthopaedic review.
- Age over 40 with a first-time dislocation and persistent weakness or limited active elevation. High prevalence of acute rotator cuff tear. Early MRI within 2–4 weeks.
- Recurrent dislocations with progressive numbness or weakness between events. Quadrilateral space syndrome, axillary nerve entrapment.
- Suspicion of seizure-related dislocation. Bilateral instability from a convulsive event. Neurology workup; rule out posterior pathology.
- Engaging Hill-Sachs lesion with a documented mechanical block to motion. Surgical referral; conservative trial wastes time and damages bony architecture further.
Refer to: A&E for acute neurovascular compromise or irreducible dislocation. Orthopaedic shoulder surgeon for surgical candidacy, recurrent dislocations, or bone-loss assessment. Neurology for persistent nerve deficit.
Right now, lie on your back and have someone slowly raise your arm out to the side and twist the hand back behind your head. If the shoulder feels like it is about to slip, that is the apprehension sign. A positive apprehension at end-range plus a clear traumatic dislocation in the past is enough to book an in-person assessment this week.
The Verdict
If your shoulder has popped out once, the choice you make in the next six weeks probably decides whether it ever pops out again.
Your shoulder is a golf ball sitting on a tee, and the soft cartilage rim around the tee is what holds the ball in place. A first dislocation rips that rim off the bone. The body cannot reliably grow that rim back through exercise alone. Each subsequent dislocation also chips a small piece off both the ball and the tee, so the joint actually gets less stable, not more, with every event you let happen.
- What this actually is: A traumatic dislocation tears the cartilage rim (the labrum) and stretches the ligaments at the front of the shoulder. The rim does not regrow. Without that rim, the dynamic muscles alone cannot reliably hold the joint in place under high-demand loading.
- The myth that won't die: That every young athlete should try a long rehab trial first. Five recent meta-analyses agree — for under-25 contact athletes, early arthroscopic Bankart repair cuts the chance of another dislocation by about four times.
- Start here: Get an MRI arthrogram if surgery is being considered. For older or low-demand patients, structured rehab over 12–24 weeks works for most. For young contact athletes, the surgical conversation should happen early — every delayed dislocation makes a future surgery harder.
Conviction
What would change this: A multi-center RCT (N≥400, age 18–30 contact athletes, first-time dislocators) with a standardized, dosed conservative rehab arm vs early arthroscopic Bankart, primary endpoint 5-year recurrence + glenohumeral OA on imaging, secondary endpoints including TSK and SI-RSI. Until that exists, early surgery for young contact athletes and conservative-first for older low-demand patients remains the appropriate stratification.
Sub-conviction breakdown
Surgical stabilization superior to conservative for young (<25) contact athletes: HIGH — five SR/MAs concordant, RR ~0.20–0.25, BESS/BOA Patient Care Pathway.
ER vs IR sling immobilization equivalence: HIGH — Zhang 2020 SR/MA of RCTs N=1042.
Recurrence drives bone loss and OA progression: HIGH — Rutgers 2022 SR; Verweij 2021 SR N=1832.
Critical bone-loss threshold ~13.5–20% for Latarjet over Bankart: MODERATE — international consensus; primary data variable.
Conservative management appropriate for >30 yo low-demand: MODERATE — pooled data thinner; based on lower baseline recurrence risk.
Specific conservative-rehab dosing (sets/reps/load by phase): LOW — DATA UNAVAILABLE in most RCTs.
Apprehension+relocation+surprise cluster diagnostic accuracy: MODERATE-HIGH — consistent across Provencher 2021 review and primary cohort data.
Criteria-based RTP reduces recurrence vs time-based: MODERATE — Kelley 2022 cohort, not RCT-validated head-to-head.
Psychological factors substantially affect RTS: HIGH — Velasquez Garcia 2023 SR/MA N=1093.
Don't want to guess what works the next time you're injured? Get The Verdict — free weekly evidence-graded protocols.
Subscribe FreeThe Full Picture — Anatomy, Diagnosis & Evidence
What's Actually Going On
The shoulder is a shallow socket (the glenoid) holding a large ball (the humeral head). Stability comes from a fibrocartilage rim called the labrum, the inferior glenohumeral ligament complex, the joint capsule, and dynamic muscle control from the rotator cuff and scapular stabilizers.
A traumatic anterior dislocation forces the humeral head forward and downward off the glenoid. This typically tears the front-lower part of the labrum off the bone (a Bankart lesion, present in 80–100% of first-time MR arthrography findings) and impacts the back of the humeral head against the front of the glenoid rim, creating a divot (a Hill-Sachs lesion, present in 65–93%). With each subsequent dislocation, more bone is lost from both surfaces — the structural problem grows worse with every event.
Once the labrum is detached and the capsule stretched, the dynamic stabilizers alone cannot reliably prevent the joint from slipping again under abduction–external rotation loading. This is why first-time dislocations in young, high-demand patients have such high recurrence rates without surgical repair.
How to Identify It
Typical presentation: "My shoulder came out — it popped back in, but ever since, when I try to throw or reach behind my head, it feels like it's about to slip again."
No single physical test rules in or out anterior instability with confidence. Use the apprehension–relocation–surprise cluster as the clinical battery; suspicious positive cluster with mechanism-consistent history → MR arthrography → 3D CT only if surgical planning needs bone-loss quantification.
| Test | Sensitivity | Specificity | Source |
|---|---|---|---|
| Apprehension test | 53–72% | 96–99% | Provencher 2021 |
| Relocation test | 30–81% | 54–92% | Provencher 2021 |
| Surprise / Anterior Release | 64–92% | 89–99% | Provencher 2021 |
| 3-test cluster (all positive) | — | ~97–99% (PPV ~93–94%) | Provencher 2021 |
| MR arthrography (labral lesions) | 88–93% | 94–98% | Provencher 2021 |
Imaging trigger: Order MR arthrography only when surgery is on the table; reserve 3D CT for glenoid bone-loss quantification approaching the critical threshold (~13.5–20%).
The Debate
Older view → Pre-2010 orthopaedic / primary-care guidance
Conservative management is the default first-line for first-time anterior dislocation in all populations.
Recent evidence → Hurley 2020, Belk 2023, Alkhatib 2022, Hu 2023, Abdel Khalik 2024 SR/MAs; BESS/BOA Patient Care Pathway
Early arthroscopic Bankart repair is first-line for young (<25), contact-sport, high-demand first-time dislocators — approximately 4-fold recurrence reduction.
Older view → Internal-rotation sling for 3–4 weeks reduces recurrence; Itoi early RCT favored ER bracing
Sling position determines recurrence risk.
Recent evidence → Zhang 2020 SR/MA of RCTs (N=1042)
No significant difference in recurrence between ER and IR immobilization. Choose by patient comfort and compliance — ER bracing has high real-world non-compliance.
Older view → Conservative management produces equivalent long-term joint outcomes to surgery
Surgery is purely an option for those wanting fastest functional return.
Recent evidence → Verweij 2021 SR/MA (N=1832)
Post-dislocation osteoarthritis prevalence is driven by recurrence count, not by surgery itself. Reducing recurrences protects the joint long-term.
Older view → Time-based return to play at 4–6 months post-stabilization
Calendar time governs RTS clearance.
Recent evidence → Kelley 2022 cohort
Criteria-based RTP testing reduced recurrence to 5% vs 22% with time-only criteria. Time alone does not capture neuromuscular and psychological readiness.
Honest Limitations
The conservative-rehab dosing gap
Surgical-vs-conservative RCTs uniformly describe their conservative arms as "supervised rehabilitation" with no sets, reps, load, or progression criteria. Chiddarwar 2023 (BJSM, N=3598) explicitly identified this as a limitation. Clinicians applying "conservative management" reproduce unknown protocols. Frame any "12-week trial" as 12 weeks of real phase-criteria-driven loading, not calendar time.
RCT populations are not the average patient
Most surgical-vs-conservative RCTs enroll younger motivated athletes presenting to specialist centers. Davey 2023 fragility index analysis shows the median FI is 2 — many headline findings rest on flipping 1–2 patient outcomes. Stratify by age, demand, and bone-loss status; the 4-fold recurrence reduction is most robust in young contact athletes and diminishes in older or lower-demand cohorts.
Psychological readiness is rarely measured
Velasquez Garcia 2023 (N=1093) shows ~34% of physically-eligible post-stabilization athletes do not return at pre-injury level. Most clinics gate RTS on physical metrics only. Add Tampa Scale of Kinesiophobia and SI-RSI screening at the late-rehab milestone; pain neuroscience education and graded exposure are appropriate when scores indicate elevated fear.
The Nuance
' width='800' height='400'/%3E%3C/svg%3E)
Conservative success rate is age- and demand-dependent. For under-25 contact athletes, only ~30–50% achieve full functional return without recurrence at 24 months. For 25–40 year-old recreational patients, ~50–80%. For over-40 low-demand patients, ~60–85% (with the caveat that cuff status often matters more than the labrum).
Surgical (arthroscopic Bankart) outcomes: ~85–92% no recurrence at 24+ months in first-time dislocators across Hurley 2020, Belk 2023, Alkhatib 2022, Hu 2023, and Abdel Khalik 2024. Latarjet ~94–96% no recurrence in recurrent or critical-bone-loss cases.
When surgery IS indicated: Young contact-sport first-time dislocator with a Bankart lesion. Recurrent dislocations (≥2) regardless of age in active patients. Glenoid bone loss approaching ~13.5–20% on 3D CT. Engaging Hill-Sachs with mechanical block. Failed structured 12–16 week conservative trial in a motivated patient. Combined posterior or SLAP labral lesion identified on MRA.
When conservative IS sufficient: Age over 30 with a low-demand activity profile and an isolated anteroinferior Bankart. Age over 40 with a first-time dislocation and no critical bone loss (with concurrent cuff workup). Atraumatic / hyperlaxity-driven instability without major structural lesion. Patient declines surgery and accepts the elevated recurrence risk with informed counsel.
The honest truth: For a young, active patient with a first-time anterior dislocation, the data is now clear that early arthroscopic Bankart repair reduces the recurrence rate roughly fourfold compared with a conservative trial. Each subsequent dislocation makes a future surgery harder by progressively damaging the bony rim. For an older, lower-demand patient, conservative management has a much better chance of working on its own. The decision is not "surgery vs no surgery" — it is "which path matches this patient's age, demands, structural injury, and values," made with imaging and informed counsel.
Sources
- Hurley ET et al. (2020). Arthroscopic Bankart repair vs conservative management for first-time traumatic anterior shoulder instability: SR/MA. Arthroscopy. PMID 32389771. N=569 across 7 studies. Recurrence 13.0% vs 52.9%, OR 0.10.
- Belk JW et al. (2023). Shoulder stabilization vs immobilization for first-time anterior dislocation: SR/MA of Level 1 RCTs. Am J Sports Med. PMID 35148222. N=126.
- Alkhatib N et al. (2022). Bankart repair vs conservative for first-time anterior dislocation: SR/MA of RCTs. J Shoulder Elbow Surg. PMID 35398165. N=348.
- Hu B et al. (2023). Arthroscopic Bankart vs conservative for first-time dislocation: SR/MA. Eur J Med Res. PMID 37501089. N=786 across 9 RCTs. Recurrence RR 0.23 (95% CI 0.16–0.34).
- Abdel Khalik H et al. (2024). Arthroscopic stabilization for first-time anterior dislocation: SR/MA. J Shoulder Elbow Surg. PMID 38430981.
- Imam MA et al. (2021). Bankart vs Latarjet for recurrent anterior instability: SR/MA. Am J Sports Med. PMID 33264030. N=3275 shoulders. Latarjet 5.7% vs Bankart 11.4% recurrence.
- Gilat R et al. (2021). Latarjet vs free bone block for anterior instability with bone loss: SR/MA. Am J Sports Med. PMID 32795174. N=4540 shoulders.
- Hurley ET et al. (2020). Remplissage for anterior instability with Hill-Sachs lesions: SR/MA. J Shoulder Elbow Surg. PMID 32650087.
- Hurley ET et al. (2022). Anterior Shoulder Instability International Consensus Part I: diagnosis, nonoperative management, Bankart repair. Arthroscopy. PMID 34332055.
- Hurley ET et al. (2022). Anterior Shoulder Instability International Consensus Part II: Latarjet, remplissage, glenoid bone-grafting. Arthroscopy. PMID 34332052.
- Provencher MT et al. (2021). Diagnosis and management of traumatic anterior shoulder instability. JAAOS. PMID 33275397.
- Zhang B et al. (2020). External vs internal rotation immobilization after shoulder dislocation: MA of RCTs. OTSR. PMID 32446811. N=1042. No significant difference in recurrence.
- Verweij LPE et al. (2021). Treatment type and post-dislocation OA prevalence: SR/MA. KSSTA. PMID 32936334. N=1832.
- Rutgers C et al. (2022). Recurrence increases prevalence of Hill-Sachs and Bankart lesions: SR/MA. KSSTA. PMID 34988633.
- Velasquez Garcia A et al. (2023). Psychological factors and RTS after stabilization: SR/MA. KSSTA. PMID 37991534. N=1093.
- Kelley TD et al. (2022). Functional rehab and criteria-based RTP after arthroscopic stabilization. Sports Health. PMID 34918564. Criteria-based 5% vs time-only 22% recurrence.
- Davey MS et al. (2023). Fragility Index of RCTs in anterior shoulder instability. Am J Sports Med. PMID 35414266. Median FI = 2.
- Chiddarwar V et al. (2023). Combined surgical + exercise interventions for primary traumatic anterior dislocation: SR/MA. Br J Sports Med. PMID 37451706. N=3598.
- Saleet J et al. (2025). Contact vs noncontact athletes after primary arthroscopic Bankart repair: SR/MA. Am J Sports Med. PMID 40574344.