Get Labs

Overview

get_labs retrieves laboratory results for a specific patient from FHIR Observation resources. An LLM/agent
should call this tool when the user asks about lab results, blood tests, panels, or specific laboratory
measurements — CBC, metabolic panel, cholesterol, glucose, hemoglobin A1c, or similar. For neutrophils/lymphocytes,
use 'Complete Blood Count' (includes all cell differentials); use full test names rather than abbreviations
(e.g. 'CRP' not 'c-reactive protein').

Required Parameters

None — all parameters are optional (though person_id is effectively required unless derivable from the token).

Optional Parameters

person_id (string)

• Default: None
• The FHIR identifier (patient ID) of the person whose lab results are to be retrieved. Extracted from the access token if not provided.

labs (list of exact lab test names from the FHIR_LABS literal)

• Default: None (all available lab results)
• Invalid names are ignored (logged, or reported via ctx.info when debug=true); if every provided name is invalid, all labs are returned instead.

start_date / end_date (date, YYYY-MM-DD)

• Default: None
• Inclusive date bounds. No bound applied if omitted.

count (integer)

• Default: the environment's observations_default_count
• Number of raw FHIR records to fetch before filtering. Minimum of 300 automatically enforced; increase to 1000–5000 for comprehensive historical searches.

output_formats (list of table | timeline | statistics | csv | markdown)

• Default: ["table", "statistics"]
• table for detailed records, statistics for aggregated insights, timeline for chronological view, csv for export, markdown for documentation. Multiple formats can be combined.

ignore_cache (boolean)

• Default: false
• If true, bypasses any cached data and fetches fresh records from the FHIR server.

debug (boolean)

• Default: false
• Enable to include additional diagnostic information in the response.

Returns

FhirRecordRetrievalOutput — a result field containing the retrieved lab data as a string, formatted per output_formats.