12-phase flow (12)

1. 1FRAME
   Frame as a LONGITUDINAL tonsillectomy-candidacy engine — NOT a per-episode antibiotic engine (ent.pharyngitis.core.v1 owns the acute episode). Two jobs: (A) decide whether the patient meets Paradise criteria with adequate documentation, with watchful-waiting bias per AAO-HNSF 2019 (Mitchell PMID 30798778); (B) identify modifying factors and red-flag mimics (PFAPA, malignancy, OSA) that change the pathway.
   advance: longitudinal scope confirmed; per-episode antibiotic decisions routed to ent.pharyngitis.core.v1
2. 2ENTRY
   Verify the documented episode count meets a Paradise threshold (year-1 ≥7, year-2 ≥5/y × 2, or year-3 ≥3/y × 3) AND each episode has the required objective documentation. Episodes without documentation do NOT count — common error in real-world practice and the most frequent reason a patient appears Paradise-positive but is not (Paradise NEJM 1984 PMID 6700642; AAO-HNSF 2019 Mitchell PMID 30798778).
   inputs: episode_count_with_documentation_year_1, documentation_quality_per_episode
   advance: Paradise threshold and documentation gates evaluated explicitly
3. 3CONTEXT
   Build the longitudinal context: 2-year and 3-year episode counts, age (affects anaesthetic / overnight-monitoring / codeine restrictions), comorbidities (bleeding risk, OSA features, cardiac, syndromic), penicillin-allergy profile. OSA features always flagged — OSA tonsillectomy is a separate AAO-HNSF 2019 pathway.
   inputs: episode_count_with_documentation_years_1_2, episode_count_with_documentation_years_1_2_3, age, red_flag_obstructive_sleep_apnea_features, penicillin_allergy, medical_comorbidities_bleeding_risk_anaesthesia_risk
   advance: longitudinal episode + age + comorbidity + OSA context captured
4. 4RED_FLAGS
   Recognise (do NOT manage here): asymmetric tonsillar enlargement / unilateral mass / B symptoms → malignancy work-up (lymphoma, head/neck SCC); acute peritonsillar abscess → ent.peritonsillar-abscess.core.v1; sepsis / Lemierre with current acute episode → id.sepsis.core.v1; airway compromise from acute episode or perioperative complication → emergent airway. Post-tonsillectomy haemorrhage is a known complication and a return-precaution.
   inputs: red_flag_tonsillar_asymmetry_or_b_symptoms
   actions: workup.airway_distress, calc.qsofa
   advance: malignancy / abscess / sepsis / airway red flags screened and routed if positive
5. 5INITIAL_WORKUP
   Confirm GAS at each documented episode (RADT, with back-up culture in children per IDSA 2012); collect documentation (date, symptom + objective criterion + treatment) into a longitudinal log. AAO-HNSF 2019 strongly discourages routine perioperative antibiotics. Pre-tonsillectomy polysomnography is required for children <2 y or with obesity / Down syndrome / craniofacial / neuromuscular / sickle / mucopolysaccharidoses, and is recommended for any child where the need for tonsillectomy is uncertain or discordant with exam (Mitchell PMID 30798778).
   inputs: rapid_strep_radt, throat_culture_back_up
   actions: panel.cbc, panel.inflammation
   advance: per-episode GAS confirmation and longitudinal log built; polysomnography ordered if OSA pathway
6. 6BRANCHING_WORKUP
   Branch on pattern: regular monthly fever + aphthous + pharyngitis + adenitis → PFAPA (consider IL-1 axis pathway / specialist immunology) — tonsillectomy can be curative; tonsillar asymmetry / B symptoms → ENT urgent referral for biopsy (lymphoma / SCC); OSA features → polysomnography + AAO-HNSF 2019 OSA tonsillectomy pathway; recurrent peritonsillar abscess → tonsillectomy modifying factor
   inputs: modifying_factor_pfapa_pattern, modifying_factor_prior_peritonsillar_abscess
   actions: workup.lymphadenopathy
   advance: PFAPA / OSA / malignancy / abscess branches assigned and routed if needed
7. 7DIFFERENTIAL
   Terminal differential: classical recurrent GAS tonsillitis meeting Paradise (treat or watchful-wait per AAO-HNSF 2019) vs Paradise-not-met / sub-threshold (watchful waiting recommended) vs PFAPA (regularity + aphthous + adenitis pivot — tonsillectomy can be curative) vs OSA / sleep-disordered breathing (snoring + apnoeas + polysomnography pivot — separate tonsillectomy pathway) vs streptococcal carrier (asymptomatic positive culture — do NOT treat or operate) vs malignancy (asymmetry / mass / B symptoms — biopsy pathway) vs infectious mononucleosis (longer fatigue + posterior nodes + splenomegaly — different pathway). The chosen pathway drives the next step.
   advance: single best pathway chosen; sub-threshold cases held for watchful waiting
8. 8RISK_STRATIFICATION
   Stratify tonsillectomy candidacy: STRONG (meets Paradise with adequate documentation AND symptomatic burden / time-off-school-work AND failed watchful-waiting OR has PFAPA / OSA / recurrent PTA) vs MODERATE (modifying factors present but Paradise not strictly met — shared decision) vs WEAK (Paradise not met, no modifying factor — AAO-HNSF 2019 RECOMMENDS WATCHFUL WAITING). Layer perioperative risk: bleeding diathesis, age <3 y or severe OSA (overnight admission required), codeine contraindication <12 y.
   inputs: modifying_factor_prior_peritonsillar_abscess, modifying_factor_multiple_antibiotic_allergies, medical_comorbidities_bleeding_risk_anaesthesia_risk
   advance: candidacy band assigned with perioperative-risk layered
9. 9TREATMENT
   Per-episode antibiotic regimen for confirmed GAS is OWNED by ent.pharyngitis.core.v1 (penicillin V / amoxicillin first-line × 10 days; benzathine penicillin G 1.2 MU IM single dose; cephalexin if non-anaphylactic penicillin allergy; clindamycin / azithromycin if severe). Symptomatic care every episode (analgesia, hydration; single-dose dexamethasone for severe odynophagia is an option). LONGITUDINAL decision: AAO-HNSF 2019 strong recommendation = WATCHFUL WAITING if Paradise not met; tonsillectomy is an OPTION (not recommendation) when met. Post-tonsillectomy: single intraoperative IV dexamethasone (AAO-HNSF 2019 strong recommendation); ibuprofen + acetaminophen for pain control (AAO-HNSF 2019 strong recommendation); NO codeine in children <12 years (AAO-HNSF 2019 strong recommendation against); NO routine perioperative antibiotics (AAO-HNSF 2019 strong recommendation against).
   inputs: penicillin_allergy, pregnancy
   advance: per-episode regimen routed to ent.pharyngitis; longitudinal candidacy decision documented; perioperative bundle ordered if proceeding
10. 10DISPOSITION
    OUTPATIENT clinic-based longitudinal management is the default. Inpatient overnight monitoring REQUIRED post-tonsillectomy for: age <3 years; severe OSA (AHI ≥10 obstructive events/h, SpO2 nadir <80%, or both); other significant comorbidity. Same-day discharge appropriate for older healthy patients. Acute peritonsillar abscess or sepsis with the current episode → admit and route OUT (Mitchell PMID 30798778).
    inputs: age, red_flag_obstructive_sleep_apnea_features
    advance: disposition documented; inpatient-monitoring criteria explicitly applied to post-op cases
11. 11MONITORING
    Pre-tonsillectomy longitudinal: track episodes, documentation quality, and patient-reported burden (school/work absence, QoL); reassess at 6-12 months — many sub-threshold patients regress to fewer episodes naturally. Post-tonsillectomy: assess primary (<24 h) and secondary (>24 h) bleeding rates per individual surgeon (AAO-HNSF 2019 requires annual measurement). Post-op pain control reassessment 24-72 h. Counsel that obstructive sleep-disordered breathing MAY persist or recur post-tonsillectomy (Mitchell PMID 30798778).
    inputs: documentation_quality_per_episode
    advance: longitudinal episode-tracking documented; post-op recovery reviewed
12. 12FOLLOWUP
    Pre-op: 6-12 monthly review with formal episode-and-documentation re-count; shared-decision conversation each visit. Post-op: review at 2-3 weeks for haemorrhage / hydration / pain (peak secondary bleed window day 5-10), at 6-8 weeks for symptom resolution + OSA re-evaluation (recurrence possible). Document primary + secondary bleeding outcome in medical record (AAO-HNSF 2019 requirement). Long-term: counsel that tonsillectomy benefit decays beyond 12 months in the moderately-affected group (Paradise Pediatrics 2002 PMID 12093941).
    inputs: documentation_quality_per_episode
    advance: structured longitudinal review + post-op bleeding-outcome documentation completed