sepsis
Enable an AI agent to recognise a suspected sepsis episode, represent its evidence and evolving severity, and support timely clinical escalation and reassessment.
Research draft, second pass
A second pass drafted this model: the structure a model of this thing needs, and what is known about it in the world. The line under this one says how the second half was obtained - researched against sources, or recalled without web access, in which case nothing here was read anywhere and every claim is a lead to verify. Unreviewed either way.
recalled by Codex without web access - no source was read
Researched by: Codex
Purpose and description
Enable an AI agent to recognise a suspected sepsis episode, represent its evidence and evolving severity, and support timely clinical escalation and reassessment.
Sepsis is life-threatening organ dysfunction caused by a dysregulated host response to infection.
It can be Assemble a time-stamped evidence account for suspected sepsis and identify missing observations.; Compare the episode with an applicable, explicitly versioned definition while preserving uncertainty.; Flag possible deterioration for urgent clinician review through the configured care pathway.; Track antimicrobial, source-control and organ-support decisions, dependencies and reassessment.; Identify contradictions between diagnostic labels, organ dysfunction evidence and alternative explanations.; Prepare a handover of unresolved infection, organ dysfunction and follow-up needs..
Distinguishing features
Sepsis requires life-threatening organ dysfunction associated with a dysregulated response to infection; an uncomplicated infection does not meet that conceptual boundary. See [WHO sepsis overview](https://www.who.int/health-topics/sepsis/sepsis).
Record evidence of infection separately from bloodstream culture results: bacteremia and sepsis are different classifications.
Fever, tachycardia or systemic inflammatory response alone do not establish sepsis; record organ dysfunction and competing explanations.
Septic shock is a subset of sepsis with profound circulatory and metabolic abnormalities; neither isolated hypotension nor elevated lactate alone establishes that classification. See [Sepsis-3](https://jamanetwork.com/journals/jama/fullarticle/2492881).
Separate diagnostic criteria from risk screening: neither qSOFA nor SOFA alone constitutes a stand-alone sepsis definition. See [Sepsis-3](https://jamanetwork.com/journals/jama/fullarticle/2492881).
Scope
+ Suspected or established infection and its relationship to acute organ dysfunction
+ Population-specific sepsis definitions and operational criteria
+ Organ dysfunction, perfusion abnormalities and septic shock status
+ Time-stamped recognition, escalation, treatment dependencies and reassessment
+ Episode resolution, persistent impairment and transition to follow-up
- Pathogen taxonomy and complete models of individual infectious diseases
- Noninfectious shock and inflammatory disorders except as differential diagnoses
- Complete models of individual organ failures
- Drug formularies, dosing algorithms and procedural specifications
- Hospital-wide infection prevention programmes and population surveillance systems
Characteristics
- Episode assessment
- suspected | clinically established | uncertain | ruled out | resolving | resolved; timestamp and assessor Separates evolving clinical judgment from a permanent disease label.
- Definition framework
- named framework, version, target population and intended use Makes adult, pediatric, neonatal and surveillance classifications interpretable.
- Infection evidence
- links to suspected anatomical source, microbiology, imaging and clinical assessment Exposes the evidence supporting the infectious trigger.
- Organ dysfunction attribution
- affected organ system linked to baseline, acute change, infection attribution and alternative causes Distinguishes acute infection-associated dysfunction from chronic impairment or another acute process.
- Organ dysfunction score
- points; instrument, component values, baseline, observation window and missingness Supports reproducible operational assessment without treating a score as the diagnosis.
- Perfusion observations
- blood pressure in mmHg, lactate in mmol/L, urine output in mL/kg/h and capillary refill in seconds; timestamps and context Supports assessment of circulatory deterioration and response.
- Septic shock assessment
- suspected | criteria met | criteria not met | indeterminate | resolved; framework and supporting observations Distinguishes urgent clinical suspicion from a completed criteria-based classification.
- Episode timing
- date-time with timezone for first concern, recognition, interventions and reassessments; estimated onset interval Separates observed events from retrospective estimates of onset.
- Active treatment context
- links to antimicrobial therapy, source control, resuscitation and organ support Allows observations and trajectory to be interpreted alongside interventions.
Also called
Where this came from
wikidata · CC0 1.0
Drafted structure
Bundle to layer to finding to question, as the second pass will find it: 6 bundles · 11 layers · 19 findings · 29 questions.
Sepsis identity and criteria Establish what qualifies as sepsis for this episode and which assessment framework applies.
A sepsis label must expose its clinical meaning, population and evidential basis.
Clinical definition
Represent the connection between infection and life-threatening organ dysfunction.
Infection-dysfunction relationship
Record infection evidence, acute dysfunction and the assessor's causal judgment separately.
- What supports infection as the cause of the acute organ dysfunction? definition
- Which observations or alternative explanations weaken that attribution? boundary
Population and operationalisation
Identify the criteria appropriate to the patient and purpose.
Applicable criteria
Record age, population exclusions and framework version. SCCM recommends Phoenix criteria for pediatric sepsis, including suspected infection and a Phoenix score of at least 2. See [SCCM sepsis definitions](https://www.sccm.org/clinical-resources/sepsis-definitions).
- Which framework applies to this patient's age, setting and clinical circumstances? boundary
- Is this classification being used for diagnosis, screening, research or surveillance, and who supplied it? provenance
Infectious trigger and attribution Characterise the suspected infection and uncertainty about its causal role.
Sepsis management depends on identifying an infectious trigger while remaining open to competing diagnoses.
Source and organism evidence
Connect anatomical source hypotheses with the evidence supporting them.
Infection evidence account
Keep suspected source, identified organism and specimen interpretation distinct, with collection times and treatment context.
- Which anatomical source is suspected, and what clinical, imaging or microbiological evidence supports it? provenance
- When were specimens obtained relative to antimicrobial treatment, and are results pending or potentially contaminated? measurement
Competing and concurrent causes
Represent other explanations for deterioration and possible coexistence with sepsis.
Differential attribution
Record competing causes such as hemorrhage, cardiogenic shock or medication effects without assuming that they exclude concurrent infection.
- Which competing or concurrent process could explain each major abnormality? boundary
- What new evidence would prompt revision of the sepsis assessment? action
Organ dysfunction and shock Represent acute physiological impairment and the evidence for septic shock.
The model must distinguish baseline impairment, current dysfunction and worsening circulatory failure.
Acute organ change
Track organ-specific abnormalities against a defensible baseline.
Baseline-adjusted dysfunction
Record respiratory, cardiovascular, renal, hepatic, coagulation and neurological observations as available, including treatment confounders and missing measurements.
- Which organ functions have changed acutely, compared with what documented or assumed baseline? measurement
- How do sedation, ventilation, chronic disease or missing data affect interpretation? boundary
Circulation and perfusion
Track hemodynamic support and perfusion evidence with the relevant shock framework.
Shock classification evidence
Adult Sepsis-3 operational criteria include vasopressors needed to maintain mean arterial pressure of at least 65 mmHg and lactate above 2 mmol/L despite adequate volume resuscitation. Record the assessment context. See [Sepsis-3](https://jamanetwork.com/journals/jama/fullarticle/2492881).
- What do blood pressure, lactate, perfusion findings and support requirements show over time? measurement
- Are the applicable shock criteria met, unmet or unassessable, and how was volume status evaluated? boundary
Recognition and care response Connect urgent clinical concern with accountable actions and review.
An episode model must support action under uncertainty and preserve the reasons for care decisions.
Recognition and escalation
Record when concern arose and how it reached the responsible clinical team.
Clinical response ownership
Sepsis requires prompt recognition and management; capture escalation, responsible clinician and reassessment plan. See [WHO sepsis overview](https://www.who.int/health-topics/sepsis/sepsis).
- When was possible sepsis recognised, by whom, and when was the responsible clinician notified? provenance
- What immediate review or escalation does the applicable care pathway require? action
Treatment decisions and dependencies
Link episode-level decisions to treatment and procedure records.
Infection and support response
Track antimicrobial decisions, source-control assessment, resuscitation and organ support, including timing, constraints and observed response.
- Which infection-directed and supportive actions were ordered and delivered, and what remains pending? action
- What findings, contraindications or goals of care justify continuing, modifying or stopping each intervention? action
Trajectory and episode closure Represent change over time and the basis for closing or transferring the episode.
Improving observations, reduced support and resolution of infection may occur at different times.
Serial reassessment
Compare successive observations alongside treatment intensity.
Response versus supported stability
Record whether apparent improvement coincides with reduced, unchanged or increased support.
- Are organ dysfunction and perfusion improving, stable or worsening across comparable observation windows? measurement
- Does the trajectory require renewed source investigation, diagnostic revision or escalation? action
Resolution and residual needs
Separate closure of the acute sepsis episode from recovery of all affected functions.
Closure evidence and handover
Record the clinician's basis for resolution or diagnostic revision and link remaining impairments and care needs to follow-up records.
- What evidence and clinical judgment support marking the episode resolved or ruled out? boundary
- Which persistent impairments, pending results and treatment obligations require handover? action
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.
A model that cannot be lined up against existing standards, identifiers and practice cannot be adopted by anyone who already uses them.
Reported evidence
Findings from the breadth pass, kept separate from the structural claims.
Check these first
Recalled without web access and unsourced; every item is a lead to verify.
- This describes the clinical syndrome, although the supplied domain code may suggest another classification; registry placement should be checked.
- The numerical criteria listed are adult Sepsis-3 criteria; pediatric, neonatal and pregnancy-specific assessment requires separate verification.
- This is recalled knowledge without source consultation; verify current guideline editions and jurisdiction-specific coding before operational use.
- Which of these check these first hold for the sense of sepsis this model covers, and on what evidence? provenance
Kinds and varieties
Recalled without web access and unsourced; every item is a lead to verify.
- Bacterial sepsis
- Fungal sepsis
- Viral sepsis
- Neonatal sepsis
- Maternal sepsis
- Septic shock, a subset with profound circulatory and cellular or metabolic abnormalities
- Which of these kinds and varieties hold for the sense of sepsis this model covers, and on what evidence? provenance
Identifiers and schemes
Recalled without web access and unsourced; every item is a lead to verify.
- ICD-10 - A40-A41 - These categories cover streptococcal and other sepsis; additional or different codes apply to some pathogens and obstetric or neonatal contexts, with national modification differences.
- Which of these identifiers and schemes hold for the sense of sepsis this model covers, and on what evidence? provenance
Standards and regulation
Recalled without web access and unsourced; every item is a lead to verify.
- Sepsis-3 international consensus definitions, developed by a task force convened by the Society of Critical Care Medicine and the European Society of Intensive Care Medicine.
- Surviving Sepsis Campaign international management guidelines, issued jointly by the Society of Critical Care Medicine and the European Society of Intensive Care Medicine.
- Which of these standards and regulation hold for the sense of sepsis this model covers, and on what evidence? provenance
Real-world use
Recalled without web access and unsourced; every item is a lead to verify.
- Recognition and escalation of care for infection-associated organ dysfunction.
- Coordination of antimicrobial treatment, infection source control and organ support.
- Hospital quality improvement and review of delays in recognition or treatment.
- Clinical trial eligibility and outcome classification.
- Public health surveillance and disease-burden estimation.
- Which of these real-world use hold for the sense of sepsis this model covers, and on what evidence? provenance
Typical measurements
Recalled without web access and unsourced; every item is a lead to verify.
- Acute change in Sequential Organ Failure Assessment score - An increase of at least 2 points operationalises organ dysfunction in adult Sepsis-3; this is a criterion, not a typical patient range. - points
- Blood lactate - Greater than 2 is part of adult Sepsis-3 septic shock criteria when accompanied by the specified vasopressor requirement despite adequate volume resuscitation. - mmol/L
- Mean arterial pressure - A vasopressor requirement to maintain at least 65 is part of adult Sepsis-3 septic shock criteria, together with elevated lactate despite adequate volume resuscitation. - mmHg
- Which of these typical measurements hold for the sense of sepsis this model covers, and on what evidence? provenance
Failure modes and hazards
Recalled without web access and unsourced; every item is a lead to verify.
- Delayed recognition because early manifestations are nonspecific or fever is absent.
- Progression to septic shock, multiple organ dysfunction and death.
- False reassurance from negative blood cultures or a low screening score.
- Delayed effective antimicrobial treatment or failure to control the infection source.
- Persistent physical, cognitive or psychological impairment after survival.
- Which of these failure modes and hazards hold for the sense of sepsis this model covers, and on what evidence? provenance
Regional variation
Recalled without web access and unsourced; every item is a lead to verify.
- Pathogen prevalence, antimicrobial resistance and common infection sources vary geographically.
- Access to diagnostics, antimicrobials, source control and intensive care affects recognition and outcomes.
- Coding rules and surveillance definitions differ between jurisdictions, limiting direct comparisons.
- Which of these regional variation hold for the sense of sepsis this model covers, and on what evidence? provenance
Neighbouring kinds and how to tell them apart
Recalled without web access and unsourced; every item is a lead to verify.
- Infection - Infection alone does not establish sepsis; sepsis requires life-threatening organ dysfunction attributable to the dysregulated response.
- Bacteremia - Bacteremia means bacteria are present in blood; it can occur without sepsis, and sepsis can occur without demonstrable bacteremia.
- Systemic inflammatory response syndrome - This physiological response pattern can result from infectious or noninfectious causes and is neither necessary nor sufficient for adult Sepsis-3.
- Septic shock - Septic shock is a subset of sepsis identified in adults by persistent vasopressor dependence and elevated lactate despite adequate volume resuscitation.
- Severe sepsis - This older category meant sepsis with organ dysfunction; Sepsis-3 incorporates organ dysfunction into sepsis itself and treats the term as redundant.
- Which of these neighbouring kinds and how to tell them apart hold for the sense of sepsis this model covers, and on what evidence? provenance
What the second pass must settle
- Which definitions and operational criteria should this catalogue adopt for neonates, pregnancy and other populations requiring specific interpretation?
- How should unknown baseline organ function and incomplete observations affect classification without silently treating missing values as normal?
- What evidence standard should distinguish suspected from clinically established sepsis when infection remains microbiologically unconfirmed?
- How should the model reconcile bedside clinical judgment with screening, research, coding and surveillance classifications that disagree?
- What explicit rules should distinguish persistent sepsis, a new episode, relapse and residual organ dysfunction after infection control?