Skip to main content

Patient Matching

Category: Healthcare Standards
Node id: patient-matching

Standards: HIPAA, GDPR, SOC2

Handles patient identity matching, MPI operations, deduplication, merging, and cross-system linking

What it does​

  • Probabilistic and deterministic patient matching
  • MPI (Master Patient Index) search with confidence scoring
  • Duplicate patient detection and deduplication
  • Record merging with configurable strategies (most_recent, most_complete)
  • Cross-system patient identity linking
  • Configurable match confidence thresholds

Use cases​

  • Patient identity resolution across hospital systems
  • Duplicate record cleanup in EHR migrations
  • Cross-facility patient linking for HIE networks
  • MPI maintenance and quality assurance

Runtime​

  • Performance tier: fast

Input ports​

NameTypeRequiredDescription
demographicsjsonNoPatient demographics for search
patientjsonNoPatient data for matching
candidatesjsonNoCandidate patients for comparison
patientsjsonNoPatient list for deduplication
master_patientjsonNoMaster patient record for merge
merge_patientsjsonNoPatient records to merge
source_patientjsonNoSource patient for linking
target_patient_idtextNoTarget patient ID for linking

Output ports​

NameTypeRequiredDescription
resultjsonNoPatient matching data (search results, match scores, duplicates, merged records, links)
has_matchbooleanNoWhether a match was found
successbooleanNoOperation success flag

Settings​

Basic​

SettingNameTypeRequiredDefaultAllowed valuesDescription
OperationoperationenumNosearchsearch, match, deduplicate, merge, linkPatient matching operation to perform
Min Confidencemin_confidencenumberNo0.8min 0; max 1Minimum match confidence threshold (0.0-1.0)

Advanced​

SettingNameTypeRequiredDefaultAllowed valuesDescription
AlgorithmalgorithmstringNoprobabilisticMatching algorithm (probabilistic, deterministic)
Duplicate Thresholdduplicate_thresholdnumberNo0.95min 0; max 1Confidence threshold for duplicate detection (0.0-1.0)
Merge Strategymerge_strategystringNomost_recentStrategy for merging records (most_recent, most_complete)
Target Systemtarget_systemstringNounknownTarget system identifier for linking

Errors​

CodeDescription
DEMOGRAPHICS_REQUIREDDemographics data required for search operation
PATIENT_REQUIREDPatient data required for match operation
PATIENTS_REQUIREDPatient list required for deduplicate operation
MERGE_DATA_REQUIREDMaster patient and merge patients required for merge
UNSUPPORTED_OPERATIONOperation not recognized

Runtime notes​

  • Result shape varies by operation: search returns matches/best_match, match returns match_results, deduplicate returns duplicate_groups, merge returns merged_patient, link returns link object
  • Confidence thresholds: >=0.95 = exact, >=0.85 = probable, >=0.70 = possible, <0.70 = unlikely
  • Match scoring weights: SSN(30) > MRN(25) > DOB(20) > last_name(15) > first_name(10) > gender(5) > phone(3) > zip(2)
  • has_match is true when at least one match meets the confidence threshold (search) or 0.9 (match)

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.