Skip to main content

RLM FHIR Analyze

Category: AI & ML
Node id: rlm-fhir-analyze
Maturity: beta
Risk tier: B

Standards: HIPAA

Single-call LLM analysis of FHIR resources/bundles: summaries, care gaps, data quality, risk, compliance. One chat completion per run — use rlm-query for recursive whole-bundle analysis

What it does​

  • FHIR R4 resource and bundle analysis via LLM
  • Clinical summary generation from FHIR data
  • Care gap identification (screenings, immunizations, adherence)
  • Data quality assessment (completeness, coding accuracy, consistency)
  • Patient risk assessment from comorbidities and risk factors
  • Compliance checking against clinical guidelines

Use cases​

  • Generating patient summaries from FHIR bundles
  • Identifying gaps in care for quality improvement programs
  • Assessing FHIR data quality before analytics or reporting
  • Risk stratification based on FHIR patient records
  • Regulatory compliance auditing of clinical data

Runtime​

  • Performance tier: variable

Input ports​

NameTypeRequiredDescription
fhir_resourcejsonNoSingle FHIR resource to analyze
fhir_bundlejsonNoFHIR bundle to analyze
querytextNoSpecific question about the FHIR data

Output ports​

NameTypeRequiredDescription
resultjsonNoFHIR analysis result containing analysis, analysis_type, and http_status
successbooleanNoWhether the operation succeeded

Settings​

Basic​

SettingNameTypeRequiredDefaultAllowed valuesDescription
LLM Endpointllm_endpointstringYesBase URL of the LLM API
ModelmodelstringNogpt-4oModel name to use
Analysis Typeanalysis_typeenumNosummarysummary, gaps, quality, risk, complianceType of FHIR analysis to perform

Auth​

SettingNameTypeRequiredDefaultAllowed valuesDescription
API Keyapi_keysecretNoAPI key for the LLM provider

Advanced​

SettingNameTypeRequiredDefaultAllowed valuesDescription
TemperaturetemperaturenumberNo0.3min 0; max 2Sampling temperature
Max Tokensmax_tokensnumberNo4096Maximum tokens to generate

Errors​

CodeDescription
FHIR_INPUT_REQUIREDEither fhir_resource or fhir_bundle input must be provided
ENDPOINT_REQUIREDllm_endpoint is not configured
LLM_REQUEST_FAILEDLLM API call failed during FHIR analysis
HTTP_ERRORLLM endpoint returned an HTTP error status

Runtime notes​

  • Provide either fhir_resource or fhir_bundle; bundle takes precedence if both are wired
  • The analysis output is free-text LLM prose, not structured FHIR
  • Large bundles are pretty-printed before sending to the LLM; watch token limits
  • The optional query input appends a focused question to the analysis prompt

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

Something unclear? Send documentation feedback by email.