The problem
An incoming sendout result may use different test and analyte codes, omit an identifier, or disagree with the patient, order, or specimen record. Those cases need review before a result can be delivered.
Science systems / 2026
ARUP-VistA Result Interface
Prototype HL7 middleware that maps sendout results, reconciles them against a simulated LIS, and sends mismatches to review.
My role
HL7 workflow design, result data modeling, reconciliation logic, and application engineering.
What I built
- HL7 parser and common result model
- Test/analyte mapping and simulated LIS reconciliation
- Exception-review console and versioned output
- Twelve synthetic message scenarios with expected outputs
The result
A documented local prototype covering result intake through simulated delivery and exception review, with twelve message scenarios and source tests for the pipeline.
Statusprototype
Evidencedocumented
Record reviewed
Implementation
How it works
The FastAPI pipeline receives an HL7 ORU R01 message, parses it into a common result model, maps test and analyte codes, and checks it against simulated LIS records. Matched results produce versioned output; duplicates, missing data, and mismatches enter review states.
Technical context and methods
The repository contains twelve synthetic HL7 fixtures, expected outputs, an operations console, and unit and integration test code. The scenarios below describe the fixture corpus. A current full test-run result and a public running service are not attached to this case study.
Evidence details
- Single numeric result
- Qualitative result
- Multi-analyte panel
- Result with comments and narrative preservation
- Corrected result with versioned transformed output
- Duplicate message detection
- Missing patient identifier
- Missing order identifier
- Unmapped test code
- Unmapped analyte
- Specimen mismatch
- Order found with a patient mismatch
- FastAPI
- HL7 ORU R01
- Pydantic
- SQLAlchemy
- Reconciliation scoring
- Audit events