12-phase flow (12)

1. 1FRAME
   Neck pain is a symptom and a triage problem, not a single diagnosis: the Bone & Joint Decade Task Force model (Guzman Spine 2008 PMID 18204387) frames it as an EPISODIC LIFETIME occurrence with variable recovery, graded I–IV (IV = major structural pathology — myelopathy, fracture, neoplasm, infection, systemic). The engine exists to NOT miss the cannot-miss Grade IV minority, apply the trauma rule correctly, and match non-pharm-first care to phenotype (only 46% of the historical literature is scientifically admissible — Carroll Spine 2008 PMID 18204397)
   inputs: pain_duration_band, arm_pain_dermatomal
   advance: Scope framed: acuity band + axial-vs-radicular-vs-myelopathic established against the trauma gate and Grade IV screen
2. 2ENTRY
   Blunt head/neck trauma (→ CCR gate); acute (<6 wk) / subacute (6–12 wk) / chronic (>12 wk) axial pain; dermatomal arm pain (radicular); myelopathic features; whiplash mechanism; inflammatory pattern; or explicit Grade IV concern (JOSPT 2017 Blanpied PMID 28666405; Guzman Spine 2008 PMID 18204387)
   inputs: pain_duration_band
   advance: Engine entered via a recognised trigger
3. 3CONTEXT
   Trauma mechanism, duration band, age (≥65 = CCR high-risk; geriatric C-spine injury ≈2× and odontoid fracture 20% vs 5% — Touger Ann Emerg Med 2002 PMID 12192352), arm-pain distribution, malignancy history, immunosuppression/IVDU, fever/constitutional symptoms, dissection risk (recent strain/manipulation), occupation, prior episodes, psychosocial yellow flags (Guzman Spine 2008 PMID 18204387; JOSPT 2017 Blanpied PMID 28666405)
   inputs: pain_duration_band, age, arm_pain_dermatomal, cancer_history, immunosuppression_or_ivdu
   advance: Trauma + duration + phenotype + risk-driver context captured
4. 4RED_FLAGS
   TRAUMA GATE FIRST: alert/stable blunt head/neck trauma → Canadian C-Spine Rule (Stiell JAMA 2001 PMID 11597285 — derivation sens 100% [95% CI 98–100], spec 42.5%, prior 1.7%; validation NEJM 2003 PMID 14695411 — CCR sens 99.4% vs NEXUS 90.7%, spec 45.1% vs 36.8%, prior 2.0%, CCR missed 1 vs NEXUS 16; CCR LR− ≈ 0.013–0.08, LR+ ≈ 1.8; AUROC 0.91 vs gestalt 0.85 — Bandiera Ann Emerg Med 2003 PMID 12944893). High-risk factor (age ≥65 / dangerous mechanism / paraesthesias) mandates imaging; low-risk factor allows ROM testing; unable to rotate 45° L+R → image. NEXUS comparator (Hoffman NEJM 2000 PMID 10891516 — sens 99.0% [98.0–99.6], spec 12.9%, prior 2.4%, LR− ≈ 0.08). CONDITIONAL-DEPENDENCY MODEL for the trauma rule (the rule's discriminating power is NOT a fixed LR — it is conditioned on patient state): (1) RULE APPLICABILITY | alertness+stability: both CCR and NEXUS are validated ONLY in alert (GCS 15), haemodynamically stable patients — in an obtunded/intoxicated/unstable patient the rule does NOT apply and its LR− is UNINFORMATIVE (do not clear clinically; image/CT) (Stiell JAMA 2001 PMID 11597285); (2) NEXUS LR− | intoxication+distracting-injury: NEXUS's "no intoxication" and "no distracting injury" criteria are themselves rule inputs — if either is present the patient is rule-positive by definition and the negative-rule LR is not computable; (3) post-test prior | age: pre-test C-spine-injury prior is age-conditioned — geriatric (≥65 y) prior ≈4.6% vs ≈2.2% non-geriatric (≈2× — Touger Ann Emerg Med 2002 PMID 12192352), and age ≥65 is itself a CCR high-risk limb forcing imaging (so a "negative" CCR is impossible at age ≥65 with any neck pain); (4) odontoid-fracture share | age: even when imaged, fracture pattern is age-conditioned (odontoid = 20% of geriatric vs 5% of non-geriatric C-spine fractures — Touger PMID 12192352), changing the post-imaging differential; (5) rule choice | structured-vs-gestalt: a structured rule out-discriminates unstructured judgment (AUROC 0.91 vs 0.85, sens 100% vs 92.2% — Bandiera PMID 12944893), so the LR profile assumed must be the structured-rule profile, not clinician gestalt. NON-TRAUMA Grade IV screen: MYELOPATHY (Hoffmann / hyperreflexia / clonus / gait-balance loss / bladder — Fehlings Global Spine J 2017 PMID 29164034 — moderate/severe DCM = surgical; insidious, frequently missed in primary care, mean diagnostic delay >2 y — Davies BMJ 2018 PMID 29472200; myelopathic-sign LR is duration-conditioned: an isolated Hoffmann sign in acute pain is far less specific than Hoffmann + clonus + gait ataxia + bladder change in chronically progressive pain), MALIGNANCY (known cancer), SPINAL INFECTION (fever + IVDU/immunosuppression/line), FRACTURE (trauma / age >65 / osteoporosis / steroid; odontoid in elderly), VERTEBRAL/CAROTID ARTERY DISSECTION (sudden severe posterior neck/occipital pain + neuro signs ± Horner after strain/manipulation; dissection causes ≈2% of all ischaemic strokes but 10–25% in adults <45 y — pain is frequently the heralding/only symptom and the dissection LR is conditioned on age <50 + recent neck strain/manipulation + thunderclap quality — Schievink NEJM 2001 PMID 11259724), INFLAMMATORY (ASAS pattern → axSpA route)
   inputs: trauma_mechanism, dangerous_mechanism_or_paresthesias, cervical_rotation_45deg, myelopathy_signs, cancer_history, immunosuppression_or_ivdu, fever_or_systemic_symptoms, dissection_features, progressive_neuro_deficit
   actions: calc.canadian_cspine, workup.fracture_triage, panel.inflammation, panel.cbc
   advance: Trauma rule applied (positive/indeterminate → CT C-spine; negative → clinically clear, no imaging) AND all Grade IV red flags screened/escalated (progressive myelopathy or progressive deficit → urgent surgical routing)
5. 5INITIAL_WORKUP
   Mostly CLINICAL — NO routine cervical imaging for rule-negative, red-flag-negative neck pain (JOSPT 2017 Blanpied PMID 28666405; Bone & Joint Decade synthesis Carroll Spine 2008 PMID 18204397). CT C-spine ONLY if a trauma rule fires. Targeted ESR/CRP + CBC (panel.inflammation, panel.cbc) ONLY if malignancy / infection / inflammatory suspicion. Cervical radiculopathy is a CLINICAL diagnosis — Wainner test-item cluster (Spurling A + cervical distraction + ipsilateral rotation <60° + ULTT-A): single-test LR+ ≈ 2–3.5, 3/4 positive LR+ ≈ 3.5, 4/4 positive LR+ ≈ 30.3 (Wainner Spine 2003 PMID 12544957); ULTT-A best single rule-OUT (high sens, LR− low). CONDITIONAL-DEPENDENCY MODEL for the Wainner cluster (the four tests are NOT conditionally independent — naïve LR-chaining over-counts): (1) cluster LR ≠ ∏(single-test LRs): the four manoeuvres all load on the same root-compression construct (shared mechanism → positive conditional correlation given disease), so the EMPIRICAL 4/4 cluster LR+ ≈ 30.3 must be used as a unit, NOT the product of the individual LR+ values (which would massively over-estimate the post-test odds); (2) Spurling LR | reference standard: Spurling specificity 0.89–1.00 but sensitivity is reference-standard-conditioned — pooled sens 0.67 vs imaging vs 0.31 vs electrodiagnosis, pooled LR+ ≈ 3.28, LR− ≈ 0.28 (Lin Am J Phys Med Rehabil 2025 PMID 39938056; Thoomes Spine J 2017 PMID 28838857; Rubinstein Eur Spine J 2006 PMID 17013656) — so the LR applied must match the verification standard used; (3) provocative-test LR | pre-test history: a positive Spurling/distraction is only diagnostically meaningful when CONSISTENT WITH the history and other findings (spectrum/verification bias inflates apparent accuracy in pure radiculopathy cohorts — Rubinstein PMID 17013656); (4) ULTT-A rule-OUT validity | symptom pattern: ULTT-A's high-sensitivity rule-out only holds for a neurodynamic/radicular pattern — a negative ULTT-A does not exclude a non-radicular axial or referred phenotype
   inputs: fever_or_systemic_symptoms, cancer_history, arm_pain_dermatomal
   actions: workup.neck_pain, panel.inflammation, panel.cbc
   advance: Clinical assessment complete; targeted labs sent only if a Grade IV trigger present; trauma CT obtained only if a rule fired
6. 6BRANCHING_WORKUP
   Imaging by red flag ONLY: CT C-spine for trauma-rule-positive/indeterminate; URGENT cervical MRI for suspected myelopathy / cord-or-root compression / spinal infection / cord metastasis (MRI also informs DCM outcome prediction — Fehlings Global Spine J 2017 PMID 29164034); XR ± CT via fracture-triage for trauma/osteoporotic/odontoid fracture; CT/MR angiography for suspected vertebral/carotid artery dissection; HLA-B27 + sacroiliac-joint MRI if ASAS inflammatory pattern → route axSpA. NO imaging for rule-negative, red-flag-negative axial or radicular neck pain (favourable natural history — Radhakrishnan Brain 1994 PMID 8186959)
   inputs: progressive_neuro_deficit, myelopathy_signs, dissection_features, inflammatory_back_pain_features
   actions: workup.fracture_triage, panel.inflammation
   advance: Targeted imaging obtained ONLY for a fired red flag; inflammatory pattern routed to axSpA workup
7. 7DIFFERENTIAL
   Terminal phenotype: (1) AXIAL mechanical neck pain (majority — no nerve-root signature, no myelopathic signs); (2) CERVICAL RADICULOPATHY (Wainner 4-test cluster 4/4 LR+ ≈ 30.3, 3/4 LR+ ≈ 3.5, ULTT-A best rule-OUT — Spine 2003 PMID 12544957; C7 most common then C6, favourable natural history 90% asymptomatic/mild at long-term, only 26% operated — Radhakrishnan Brain 1994 PMID 8186959; recent-onset radicular pain improves substantially even with wait-and-see — arm-pain VAS −19 mm over 6 wk, with collar/physiotherapy adding ≈−12 mm vs wait-and-see, Kuijper BMJ 2009 PMID 19812130 — supports conservative-first); (3) CERVICAL MYELOPATHY (cord-level signs — escalate, NOT axial management — Fehlings Global Spine J 2017 PMID 29164034; insidious onset + diagnostic delay common — Davies BMJ 2018 PMID 29472200); (4) WHIPLASH-ASSOCIATED DISORDER (Quebec WAD grade 0–IV; ≈50% still report neck pain at 1 y, greater initial pain/symptoms/disability + passive coping + depressed mood + fear-of-movement predict slower/less complete recovery — Carroll Spine 2008 PMID 18204405; baseline depression/pain ratio >1 → OR 2.0 — Griggs PMID 25322732); (5) REFERRED / co-existing MSK look-alike (cardiac, ENT/pharyngeal, TMJ, apical lung — non-spinal; also shoulder-girdle pain mimicking C5–C6 radicular pain → msk.shoulder-pain.core.v1; rotator-cuff / cervico-scapular tendinopathy → msk.tendinopathy.core.v1; cervical facet/uncovertebral spondylotic osteoarthritis sharing the chronic axial phenotype → msk.osteoarthritis.core.v1; widespread pain + ≥3-month duration + ACR-2016 criteria favouring central sensitisation → msk.fibromyalgia.core.v1); (6) INFLAMMATORY → route rheum.axial-spondyloarthritis.core.v1; (7) serious-secondary (escalate). MECE pivot: cord signs win over root signs win over axial; central-sensitisation widespread pattern is a co-existence flag, not a substitute for excluding Grade IV
   inputs: arm_pain_dermatomal, myelopathy_signs, inflammatory_back_pain_features
   advance: Phenotype assigned; myelopathy / inflammatory / serious-secondary routed; axial / radicular / WAD retained for matched non-pharm-first care
8. 8RISK_STRATIFICATION
   Canadian C-Spine Rule (calc.canadian_cspine) drives the trauma-imaging decision (sens 99.4–100%, the high-sensitivity rule-out — Stiell PMID 11597285/14695411). Neck Disability Index (calc.neck_disability_index) for subacute/chronic functional baseline + serial tracking (JOSPT 2017 Blanpied PMID 28666405). Whiplash: Quebec WAD grade + psychosocial yellow flags — initial Bournemouth depression/pain ratio >1 → OR 2.0 (P=0.02) for poor mid-term outcome (Griggs J Back Musculoskelet Rehabil 2015 PMID 25322732); symptoms plateau ~6 weeks then gradual improvement
   inputs: neck_disability_index, wad_grade_and_yellow_flags, trauma_mechanism
   actions: calc.canadian_cspine, calc.neck_disability_index
   advance: Trauma imaging decision made via CCR; NDI baseline recorded for subacute/chronic; WAD grade + yellow-flag chronicity risk documented
9. 9TREATMENT
   NON-PHARM FIRST. Stay-active + reassurance/education + ACTIVE EXERCISE first-line across acute/subacute/chronic (JOSPT 2017 Blanpied PMID 28666405; chronic NP cervico-scapulothoracic strengthening long-term pain+function SMD −0.45 [95% CI −0.72 to −0.18], stabilisation SMD −14.90 [−22.40 to −7.39] — Gross Cochrane Man Ther 2016 PMID 27317503); avoid prolonged collar/immobilisation. PHARM SECOND: NSAID first-line with the drug × comorbidity matrix (eGFR/CVD/PUD/pregnancy — shared with msk.low-back-pain.core.v1); short skeletal-muscle-relaxant course for painful spasm; duloxetine for chronic neuropathic-predominant pain; AVOID routine opioids and routine imaging; gabapentinoid not for radicular pain by shared mechanical-spine doctrine. Cervical radiculopathy is MOSTLY SELF-LIMITED — conservative first (90% asymptomatic/mild at long-term, only 26% operated — Radhakrishnan Brain 1994 PMID 8186959). Whiplash: early activation + reassurance, address psychosocial yellow flags (Griggs PMID 25322732). INTERVENTIONAL RESTRAINT: do not escalate to disc-directed minimally invasive procedures (nucleoplasty) outside trials — insufficient evidence (de Rooij Cochrane 2025 PMID 41293993; pain MD −22.71 but NO NDI/QoL benefit). SURGERY: PROGRESSIVE MYELOPATHY = URGENT surgical decompression (moderate/severe DCM — Fehlings Global Spine J 2017 PMID 29164034); progressive/severe motor deficit or refractory radiculopathy/stenosis after adequate conservative care → spine surgery referral; structured post-op physio + education (Tederko Clin Rehabil 2019 PMID 30458634)
   inputs: pain_duration_band, renal_hepatic_cv_gi_comorbidity, current_analgesics_and_opioids, pregnancy_status, myelopathy_signs
   advance: Non-pharm core delivered + matched care set; pharm escalated only on the evidence ladder; URGENT surgical referral made if progressive myelopathy / deficit
10. 10DISPOSITION
    Almost all neck pain is managed OUTPATIENT. ED → admit / urgent surgical referral ONLY for: trauma-rule-positive cervical injury, PROGRESSIVE CERVICAL MYELOPATHY (moderate/severe DCM — Fehlings PMID 29164034), progressive/severe neurological deficit, suspected spinal infection, unstable/neoplastic or odontoid fracture, or vertebral/carotid artery dissection (JOSPT 2017 Blanpied PMID 28666405; Stiell PMID 11597285)
    inputs: myelopathy_signs, progressive_neuro_deficit, trauma_mechanism
    advance: Level of care set: outpatient matched care, or emergency escalation for a fired trauma rule / Grade IV red flag
11. 11MONITORING
    Re-screen red flags at each visit (new/progressive myelopathic signs, deficit, systemic features, dissection symptoms). Reassess function with the Neck Disability Index and re-evaluate WAD trajectory (whiplash plateau expected ~6 weeks then gradual improvement — Griggs PMID 25322732). Expect favourable course for axial pain and radiculopathy (Radhakrishnan Brain 1994 PMID 8186959); non-improvement or any new myelopathic feature → re-triage and urgent MRI/surgical referral. Monitor NSAID/SMR/opioid safety; opioid-restraint as a quality indicator (shared with msk.low-back-pain.core.v1)
    inputs: neck_disability_index, myelopathy_signs, progressive_neuro_deficit
    actions: calc.neck_disability_index
    advance: Trajectory tracked; recovery on schedule OR re-triaged for persistence / new myelopathic features
12. 12FOLLOWUP
    Self-management, activity maintenance, neck/scapulothoracic exercise-programme adherence (Gross Cochrane PMID 27317503), ergonomic and work-participation counselling (JOSPT 2017 Blanpied PMID 28666405). Episodic recurrence is expected (Bone & Joint Decade model — Guzman Spine 2008 PMID 18204387). Persistent or worsening neck pain → re-evaluate phenotype, re-stratify, consider specialist referral; reinforce against opioid escalation, unnecessary imaging, and non-evidence-based interventional procedures (de Rooij Cochrane 2025 PMID 41293993)
    advance: Long-term self-management + exercise + work-participation plan in place with return precautions (any myelopathic feature / progressive deficit) taught