HL7 for master files and MFN

Master files are boring until every order, result, and bill depends on them.

HL7 master file feeds use MFN notifications to synchronize dictionaries and reference data: providers, locations, tests, charge codes, departments, devices, and other shared tables. They are easy to underestimate because they are not usually patient-facing. Then an order fails because the test code was never loaded, or a provider cannot be matched because the practitioner master is stale.

The master-file envelope is usually MFI for the file being updated and MFE for the record-level event. The detail segment depends on the file: staff/practitioner data may use STF and PRA, locations may use location segments, and service dictionaries may use different detail structures.

MSH|^~\&|HR|CITYHOSP|EHR|CITYHOSP|20260715143000||MFN^M02^MFN_M02|MFN000501|P|2.5.1 MFI|PRA^Practitioner Master File^HL70175||UPD|||AL MFE|MAD|PRA7788|20260715143000|PRAC12345^CITYHOSP^L STF|PRAC12345|PRAC12345^^^CITYHOSP^PN||Careful^Clara^^^^^L|P|F|19760202|A|20200101 PRA|PRAC12345|MD^Doctor of Medicine^HL70360|MED^Internal Medicine^LOCAL|LIC123456^^^NZMC^LN

This is synthetic sample data for learning and testing. Open it in HL7 Soup Web before mapping it so the segment groups, repeated fields, and coded values are visible.

Record-Level Event Meaning Matters

A master file update is not simply "replace this row." MFE carries record-level intent such as add, update, delete, deactivate, or local variants. Receivers need to know whether an omitted field clears the value, leaves the old value alone, or is not part of that profile.

A safe MFN receiver logs the master file identifier, record key, event code, effective time, and result. It also sends or records MFK-style acknowledgments when the implementation requires them.

  • Agree stable keys before content mapping.
  • Keep code system and assigning authority with every identifier.
  • Test deactivation and correction, not just adds.

Why Master Files Affect Every Other Feed

Order interfaces need service dictionaries. Result interfaces need observation dictionaries. Scheduling needs resources and locations. Billing needs charge codes. Provider feeds need identifiers that match orders, results, documents, and ADT. Bad master data makes every downstream interface look unreliable even when the messages are syntactically fine.

A Practical Integration Soup Workflow

In Integration Soup, route MFN feeds by MFI-1 and record-level event. Keep a visible log of the source key, destination key, event type, and whether the receiver accepted the update. If a local system needs CSV or database output, use transformers to make the mapping explicit.

The transformers tutorial is useful for mapping source records to the target table shape, and the HL7 to CSV tutorial is a good starting point for dictionary extracts.

The Test Pack I Would Ask For

Ask for add, update, deactivate, reactivate if supported, delete/correction if supported, a record with multiple identifiers, a record with an inactive date, and an invalid code that should be rejected.