Clinical Commander

All dossiers
id.sepsis.core.v1

Sepsis / septic shock

infectious_diseaseacuteadultacuteinpatient
Start encounter →Print handoutPRODUCTION

SSC 2026 emphasises individualised fluid strategy (CLOVERS), prolonged beta-lactam infusion, balanced crystalloid preference (SMART/BaSICS/PLUS), and post-sepsis goals-of-care. Pediatric sepsis (Phoenix 2024 + SCCM peds 2020) routed via sibling id.sepsis.peds.v1 dossier (PRODUCTION). PRODUCTION blockers: no end-to-end protocol-runner test for Hour-1 bundle yet; RxCUI verification pending. Deepened 2026-05-12 (B.5 ID/CC lane): evidence.pmids expanded from 11 to 12 verified-landmark RCT/guideline anchors. Deepened 2026-05-14 (shard-5-obped-id depth-pass-1): added co-located _briefs/id.sepsis.core.v1.md + _research-bundles/id.sepsis.core.v1.md. Added outpatient post-sepsis setting playbook (PICS screen at 1/3 mo, vaccination review per ACIP 2024, medication reconciliation incl. steroid taper + SUP-PPI deprescribe, cognitive/functional rehab, mental-health surveillance). Added severity triggers: multi_organ_dysfunction_onset (Vincent SOFA, life-threatening) and antibiotic_timing_miss (Kumar 2006, severe — Hour-1 bundle adherence). Appended PMIDs (Seymour qSOFA derivation, ProCESS 24635773, Kumar 16625125), bringing evidence.pmids from 12 to 15. 2026-05-22 citation remediation (PubMed-live-verified): replaced 6 mis-attributed PMIDs — Seymour qSOFA 26903337->26903335, OVATION-65 32320003->32049269, CLOVERS 36815260->36688507, ANDROMEDA-SHOCK 30772910->30772908, PLUS 34375395->35041780, ADRENAL 29347425->29347874 (predecessors resolved to a COVID-19 case series, lignin-extraction chemistry, IBD drainage report, von-Willebrand-factor review, and a physics paper). All 12 regimen RxCUIs reverse-verified correct in RxNav; no RxCUI changes. Phenotype matrix (bacterial/viral/fungal × CAS/nosocomial × source-controlled/uncontrolled × immunocompetent/compromised × shock-present/absent) is encoded indirectly via severity_triggers (neutropenic_fever_phenotype, crbsi_suspected, antibiotic_timing_miss, multi_organ_dysfunction_onset) and via workups (workup.crbsi, workup.candidemia, workup.fuo, workup.aki) and via per-setting playbook drug logic. First-class TS field for phenotype matrix is schema-blocked — see docs/framework-audit/shard-5-obped-id-state.md Schema-blocked queue. Bayesian linkage (pre-test priors per source from ProCESS Yealy NEJM 2014; LRs for qSOFA/lactate/NEWS2/cap-refill; T_treat ≈ 20% and T_test ≈ 5% post-test bacterial-infection thresholds; routing edges to pulm.cap.core.v1, pulm.hap.core.v1, id.pyelonephritis.core.v1, id.crbsi.core.v1, id.candidemia.core.v1, neph.aki.core.v1, pulm.ards.core.v1) is documented in the co-located _research-bundles/id.sepsis.core.v1.md. ROS/DDx LR seed data audited by npm run audit:ros-ddx-coverage (cross-cutting; not touched by this shard). Prehospital recognition is currently encoded via transitions[].admit (outpatient → ed); a first-class "prehospital" DossierSetting value is schema-blocked. DEPTH-PASS-2 2026-05-18 (CL-1B id/gi, golden-template dossier): (1) PATIENT-SAFETY RxCUI correction via live RxNav reverse-lookup — norepinephrine 7980→7512 (7980=penicillin G), phenylephrine 4450→8163 (4450=fluconazole), meropenem 1665005→29561 (=ceftriaxone), vancomycin 477391→11124 (=levofloxacin), daptomycin 253014→22299 (=etanercept), metronidazole 6851→6922 (=methotrexate), micafungin 121191→325887 (=rituximab), zoster 2034776→1986820 (invalid→VZV gE recombinant); PCV20 1659280 (=ceftazidime) + influenza 1656584 (invalid) rxcui omitted (vaccines lack clean RxNorm ingredient CUI — omit-rather-than-fabricate). Confirmed-correct: vasopressin 11149, epinephrine 3992, hydrocortisone 5492, pip-tazo 74169, cefepime 20481. (2) Co-located _design-brief.md + _research-bundle.md per §5.5 items 1+2 (design_brief pointer repointed from problem-package path). (3) Bayesian §5.5.2 seed trio authored at prisma/seed/ros-and-ddx/id.sepsis.core.v1.{differentials,ros,finding-lrs}.ts — 9 shock/SIRS-mimic differentials w/ cohort-anchored priors (ProCESS/ARISE/ProMISe), 16 ROS, ≥8 LR+/≥8 LR− (qSOFA/lactate/NEWS2/cap-refill/procalcitonin per-band continuous LRs), conditional dependencies (lactate|shock, qSOFA-vs-SOFA non-independence), T_test≈5%/T_treat≈20% bacterial-infection thresholds, cross-routing edges to id.pyelonephritis/id.crbsi/id.candidemia/neph.aki/allergy.anaphylaxis/cardio.cardiogenic-shock. ros-ddx-coverage gate env-blocked (no .env/DB) → structural-substitute verification per state file Gate strategy.

Entry points (5)

  • symptom
    Suspected infection + organ dysfunction (qSOFA ≥2 / NEWS2 ≥5 / SIRS; Sepsis-3 Singer JAMA 2016)
    sepsis_recognition_trigger
  • vital_abnormality
    Hypotension with suspected infection (Sepsis-3 Singer JAMA 2016)
    hypotension_with_infection
  • lab_abnormality
    Lactate >2 with infection (SSC 2026)
    lactate_elevated
  • lab_abnormality
    Positive blood culture (SSC 2026)
    positive_blood_culture
  • symptom
    Altered mental status with fever / hypothermia (qSOFA component; Sepsis-3 Singer JAMA 2016)
    altered_mental_status_with_fever

Required inputs (16)

  • sbprequired
    vital • used at CONTEXT
    MAP ≥65 mmHg target (SSC 2026; conditional 60-65 in adults ≥65)
  • maprequired
    vital • used at TREATMENT
    Direct vasopressor titration target (SSC 2026 MAP ≥65 strong recommendation)
  • hrrequired
    vital • used at CONTEXT
    Tachycardia is a SIRS criterion + drives fluid responsiveness assessment (SSC 2021)
  • rrrequired
    vital • used at CONTEXT
    Tachypnea is qSOFA + SIRS criterion (Sepsis-3 Singer JAMA 2016)
  • temperaturerequired
    vital • used at CONTEXT
    Fever / hypothermia is SIRS criterion; hypothermia carries higher mortality (SSC 2021)
  • spo2required
    vital • used at CONTEXT
    Respiratory failure is most common organ dysfunction (Sepsis-3 SOFA respiratory component)
  • gcsrequired
    symptom • used at RED_FLAGS
    qSOFA mental status component; cap-refill / encephalopathy (Sepsis-3 Singer JAMA 2016)
  • lactaterequired
    lab • used at INITIAL_WORKUP
    Hour-1 bundle marker + lactate clearance trend (SSC)
  • wbcrequired
    lab • used at INITIAL_WORKUP
    SIRS criterion + trend (SSC 2021)
  • creatininerequired
    lab • used at INITIAL_WORKUP
    AKI staging (KDIGO 2026) + antibiotic dosing
  • bilirubinrequired
    lab • used at INITIAL_WORKUP
    SOFA hepatic component (Sepsis-3 Singer JAMA 2016)
  • plateletsrequired
    lab • used at INITIAL_WORKUP
    SOFA coag + DIC workup trigger (Sepsis-3 Singer JAMA 2016)
  • blood_culturerequired
    lab • used at INITIAL_WORKUP
    SSC: cultures BEFORE antibiotics if no delay
  • procalcitonin
    lab • used at MONITORING
    De-escalation guidance only; not for initial decision (SSC 2026 conditional recommendation)
  • recent_abx_or_hospitalizationrequired
    history • used at CONTEXT
    MDRO risk; informs empiric breadth (SSC 2026 strong recommendation)
  • immunocompromiserequired
    history • used at CONTEXT
    Neutropenic / fungal / atypical pathogen risk (IDSA 2024 febrile neutropenia)

12-phase flow (12)

  1. 1FRAME
    Adult sepsis / septic shock per Sepsis-3 (SOFA ≥2 + suspected infection); pediatric sepsis covered by sibling engine
    advance: Sepsis-3 criteria met
  2. 2ENTRY
    Recognise via qSOFA / NEWS2 / SIRS + suspected infection. SSC 2026 retains qSOFA only as risk stratification, not screening
    inputs: gcs
    actions: calc.qsofa, calc.sirs
    advance: sepsis recognized
  3. 3CONTEXT
    Source clues, comorbidities, recent abx / hospitalization, immune status, code status, lines / devices (SSC 2026)
    inputs: hr, rr, temperature, spo2, recent_abx_or_hospitalization, immunocompromise
    advance: context complete
  4. 4RED_FLAGS
    Hypotension, lactate >4, AMS, hypoxia, oligo-anuria, DIC features → Hour-1 bundle + ICU (SSC 2026 strong recommendation)
    inputs: sbp, lactate, gcs
    actions: protocol.septic_shock
    advance: red flags actioned
  5. 5INITIAL_WORKUP
    Hour-1 bundle (SSC 2026 strong): lactate, blood cultures × 2 (before abx if feasible), broad-spectrum abx within 1h (Kumar 2006), IV crystalloid 30 mL/kg if hypoperfusion, source-directed cultures + imaging, CBC, BMP, LFT, coag, ABG
    inputs: lactate, wbc, creatinine, bilirubin, platelets, blood_culture
    actions: panel.cbc, panel.renal, panel.lft, panel.coag, panel.abg, panel.inflammation
    advance: Hour-1 bundle complete
  6. 6BRANCHING_WORKUP
    Source identification (SSC 2026): CXR/CT chest (pulm), UA + culture (urinary), CT abd/pelvis (abdo), wound culture / surgical eval (SSTI), LP (CNS), echo (endocarditis), de Maeseneer aspiration if line-related
    actions: workup.fuo, workup.crbsi
    advance: source identified or empirically covered
  7. 7DIFFERENTIAL
    Anaphylaxis, hypovolemic, cardiogenic, obstructive, adrenal insufficiency, thyroid storm, drug reaction (DRESS), DKA, severe pancreatitis SIRS (Sepsis-3 Singer JAMA 2016 differential)
    advance: mimics excluded or co-managed
  8. 8RISK_STRATIFICATION
    SOFA (Sepsis-3 Singer JAMA 2016) / APACHE-II / NEWS2; ICU triage; mortality estimate; vasopressor + organ-support trajectory
    inputs: sbp, spo2, gcs
    actions: calc.sofa
    advance: severity + ICU need set
  9. 9TREATMENT
    Broad-spectrum empiric within 1h (source-tailored, MDRO-aware); SSC 2026 prolonged beta-lactam infusion after loading dose; balanced crystalloids 30 mL/kg with dynamic responsiveness reassessment (CLOVERS individualised); norepinephrine first-line vasopressor → vasopressin → epinephrine → adjunct phenylephrine; hydrocortisone 200 mg/d if ongoing pressor (STRESS-L); blood glucose <180; LMWH VTE prophylaxis; PPI per SUP-ICU; lung-protective ventilation if intubated; renal replacement therapy per indications; source control within 6-12h if surgically amenable
    inputs: map, creatinine, lactate
    advance: antibiotics + fluids + vasopressor target met; source control booked
  10. 10DISPOSITION
    ICU for shock / vasopressor / mechanical ventilation / refractory acidosis; otherwise step-down with sepsis monitoring orderset (SSC 2026)
    inputs: sbp, spo2
    advance: level of care set
  11. 11MONITORING
    Lactate q2-4h until normalised, MAP continuous, UOP ≥0.5 mL/kg/h, vasopressor dose trajectory, daily SOFA (Sepsis-3), procalcitonin q48-72h for de-escalation, DDx review at 24-72h (SSC 2026)
    inputs: lactate, creatinine
    actions: panel.renal, panel.cbc
    advance: response confirmed; antibiotics narrowed by culture
  12. 12FOLLOWUP
    Post-sepsis goals-of-care discussion (SSC 2026 emphasis); PICS screen at 1-3 months; cognitive / functional rehab; review and de-escalate abx; outpatient ID f/u for endocarditis / osteomyelitis when present
    advance: post-discharge plan documented