Zum Hauptinhalt springen

Clinical Reasoning

Category: Healthcare Standards
Node id: reasoning

Standards: HIPAA, GDPR, SOC2

Supports clinical reasoning including differential diagnosis generation, evidence assessment, reasoning chain construction, recommendation synthesis, and hypothesis validation

What it does​

  • Differential diagnosis generation from clinical findings
  • Evidence strength assessment for hypotheses
  • Clinical reasoning chain construction (complaint to conclusion)
  • Treatment recommendation synthesis with personalization
  • Hypothesis validation against clinical evidence
  • Confidence scoring for clinical judgments

Use cases​

  • Structured clinical reasoning for complex presentations
  • Differential diagnosis generation in emergency settings
  • Evidence-based treatment recommendation synthesis
  • Clinical hypothesis testing during diagnostic workup

Runtime​

  • Performance tier: fast

Input ports​

NameTypeRequiredDescription
findingsjsonNoClinical findings array (for differential)
patientjsonNoPatient demographics and context
hypothesisjsonNoHypothesis name (string) or object (for assess/validate)
evidencejsonNoClinical evidence items
complainttextNoPresenting complaint (for reasoning chain)
conclusiontextNoConclusion for reasoning chain
diagnosistextNoDiagnosis (for recommendation synthesis)
patient_factorsjsonNoPatient-specific factors
contextjsonNoClinical context data

Output ports​

NameTypeRequiredDescription
resultjsonNoReasoning data (hypotheses, evidence assessment, reasoning chain, recommendations, validation)
confidencetextNoConfidence level (low/medium/high)
successbooleanNoOperation success flag

Settings​

Basic​

SettingNameTypeRequiredDefaultAllowed valuesDescription
OperationoperationenumNodifferentialdifferential, assess, reason, synthesize, validateReasoning operation to perform

Advanced​

SettingNameTypeRequiredDefaultAllowed valuesDescription
Top Ntop_nnumberNo5min 1; max 20Number of top hypotheses to return

Errors​

CodeDescription
FINDINGS_REQUIREDClinical findings array required for differential
HYPOTHESIS_REQUIREDHypothesis required for assess/validate operations
COMPLAINT_REQUIREDPresenting complaint required for reasoning chain
DIAGNOSIS_REQUIREDDiagnosis required for recommendation synthesis
UNSUPPORTED_OPERATIONOperation not recognized

Runtime notes​

  • Result shape varies by operation: differential returns hypotheses/top_hypothesis; assess returns evidence counts/strength; reason returns chain steps; synthesize returns recommendations; validate returns is_valid/strengths/weaknesses
  • Hypotheses are ranked by probability (highest first) and capped by top_n config
  • The confidence output is a top-level string (low/medium/high) derived from the primary result
  • Findings input for differential must be an array of objects with a 'description' field

See Build any workflow for how to put nodes together and Ports, types and data flow for how port types connect.

Etwas unklar? Senden Sie Feedback zur Dokumentation per E-Mail..