Surgery Domain Ontology

FLOW ALGEBRA
TITLE: Surgery Domain Ontology
DOMAIN: Domain Ontologies / Medical Domains / Surgery
SOURCE: https://kirethresearch.org/booksai-core/ontologies/domain-ontologies/medical-domains/surgery-domain-ontology.html
VERSION: 1.0
AUTHORS: Mark Anthony Higgins — Sky (Recursive AI)
STATUS: Companion formalisation; source-page specification not independently verified.

flow "SurgeryDomainOntology" v1.0 {
  ontology:
    id: "surgery-domain-ontology"
    parent: "medical-domains"
    purpose: "Represent surgical procedures, clinical indications, anatomical targets, teams, safety checks, observations, complications and outcomes with evidence and temporal context."

  types:
    Patient: {id: string}
    Encounter: {id: string, patient_id: string, timestamp: datetime?, setting: string?}
    Procedure: {id: string, name: string, code: string?, code_system: string?, status: enum[planned, in_progress, completed, cancelled, unknown]}
    AnatomicalStructure: {id: string, name: string, terminology_id: string?}
    ClinicalCondition: {id: string, name: string, terminology_id: string?}
    Practitioner: {id: string, role: string, credential_source: string?}
    Device: {id: string, kind: string, identifier: string?}
    Intervention: {id: string, kind: string, timestamp: datetime?}
    Observation: {id: string, kind: string, value: string, unit: string?, timestamp: datetime?, provenance: string}
    Outcome: {id: string, kind: string, observed_at: datetime?, follow_up_window: string?, provenance: string}
    Complication: {id: string, kind: string, severity: string?, observed_at: datetime?, provenance: string}
    SafetyCheck: {id: string, kind: string, result: enum[pass, fail, unknown], timestamp: datetime?, provenance: string}
    Evidence: {id: string, source: string, collected_at: datetime?, reliability: number?}
    Relationship: {id: string, subject: string, predicate: string, object: string, context: string?, confidence: number, evidence: string[]}

  predicates:
    indicated_for(procedure: Procedure, condition: ClinicalCondition)
    targets(procedure: Procedure, structure: AnatomicalStructure)
    performed_during(procedure: Procedure, encounter: Encounter)
    performed_by(procedure: Procedure, practitioner: Practitioner)
    uses_device(procedure: Procedure, device: Device)
    includes_intervention(procedure: Procedure, intervention: Intervention)
    has_observation(encounter: Encounter, observation: Observation)
    has_safety_check(procedure: Procedure, check: SafetyCheck)
    has_outcome(procedure: Procedure, outcome: Outcome)
    has_complication(procedure: Procedure, complication: Complication)
    supported_by(relationship: Relationship, evidence: Evidence)

  constraints:
    - "All relationship endpoints must resolve to declared, correctly typed entities."
    - "Confidence and evidence reliability, when present, must be within [0, 1]."
    - "Distinguish planned, performed and completed procedures; do not infer completion from scheduling."
    - "Preserve patient and encounter boundaries; do not merge records on name similarity."
    - "Observations with numeric quantities require units and measurement context."
    - "Clinical timestamps must preserve timezone or explicit temporal reference when known."
    - "Do not infer surgical indication, effectiveness or causation from co-occurrence alone."
    - "Absence of recorded complications does not establish absence of complications."
    - "Access to patient-linked data requires applicable authorisation and privacy controls."

  inference:
    rule procedure_encounter_link:
      if performed_during(p,e) and valid(p) and valid(e)
      then encounter_contains_procedure(e,p)
      confidence: "inherit supporting evidence; preserve procedure status"
    rule anatomical_context:
      if targets(p,a) and performed_during(p,e)
      then anatomical_context_of(e,a)
      confidence: "derived context only; not evidence of pathology"
    rule safety_check_review:
      if has_safety_check(p,c) and c.result == fail
      then requires_safety_review(p,c)
      confidence: "retain source and check timestamp"
    rule outcome_attribution_guard:
      if has_outcome(p,o)
      then associated_outcome(p,o)
      confidence: "association only; causal attribution requires additional evidence"

  SEMANTIC_EQUIVALENCE:
    - "Map accepted clinical codes using explicitly versioned terminology systems."
    - "Distinguish synonyms, procedural variants, broader classes and exact identity."
    - "Preserve laterality, anatomical site, technique and procedural status."

  CONFIDENCE:
    range: [0,1]
    factors: [documentation_quality, evidence_reliability, temporal_consistency, coding_specificity, independent_confirmation]
    policy: "Do not increase certainty merely through inference chaining or repeated dependent records."

  VALIDATION:
    - "Check entity types, identifier resolution, temporal ordering and terminology versions."
    - "Verify procedure status and distinguish indication, observation and confirmed outcome."
    - "Flag missing units, conflicting laterality, inconsistent anatomical sites and unresolved evidence."
    - "Preserve uncertainty and conflicting findings for qualified review."

  FLOW_HASH:
    algorithm: "SHA-256"
    input: "canonicalised ontology version, typed assertions, rules and provenance"
    note: "Compute hash at execution; no hash value is claimed here."

  UPDATE_MODEL:
    strategy: "versioned, provenance-preserving incremental update"
    steps: [ingest, normalise, validate, compare, reconcile, recompute_inferences, publish]
    rollback: true

  AI_CONSUMPTION:
    representation: "typed, evidence-backed clinical knowledge graph"
    outputs: [procedure_links, anatomical_links, clinical_context, safety_flags, uncertainty_annotations, provenance_records]
    safeguards: [privacy_by_design, no_unsupported_diagnoses, no_unsupported_causal_claims, clinician_review_for_clinical_use]
}

END FLOW ALGEBRA