← Home
Topics covered
- Structure — case / specimen / block / slide data model; accession numbering.
- Interfaces — HL7 messaging, Kris report exports, scanner integration endpoints.
- Data pulls for research — how to request a de-identified cohort, which fields are available, which are unreliable.
- Quality signals — turnaround, amendment rates, missing-field rates.
- Interop — bridges to QuPath / Sectra and the analytics stack.
Research use
- All research data pulls go through the de-identification pipeline before they touch a repo. The research repos consume cleaned tables only.
- The Text Analysis pipeline is the canonical extractor for narrative report content coming out of the LIS.
- The Quality Research workstream consumes LIS timing data.
Pitfalls
- Field reliability varies. Some fields are populated inconsistently depending on subspecialty practice. The LIS repo documents which fields are reliable for research; don’t assume a field is trustworthy just because it’s present.
- Timestamp interpretation. The LIS records several timestamps (received, grossed, signed-out, amended). Each means something specific — use the documented definition, don’t guess.
- Accession numbers are identifying. Even without names, patterns of accession numbers can be identifying. Research exports use hashed accession IDs exclusively.