Consultations from Outside Labs
Characterizing referred cases: case mix, turnaround, and diagnostic outcomes
What this project is
Cases referred in from outside laboratories form a distinctive workload: harder cases, often with incomplete information, slower turnaround, and higher consult-to-consult variability. This project characterizes that workload and its quality signals.
Objectives
- Describe the case mix of external consults by organ, specimen type, and originating institution type.
- Measure turnaround from receipt to final report.
- Rate of agreement vs. disagreement with the referring diagnosis.
- Identify the subset of cases where the outside lab’s material (slides / blocks / IHC) is insufficient and requires re-requesting.
Status
| Stream | Status |
|---|---|
| Retrospective cohort identification | Complete |
| Metadata cleaning | In progress |
| Agreement coding rubric | Drafted |
| Analysis | Planning |
Data
- Pulled from the consultation log in
konsultasyon, filtered tosource = external. - Kris reports for the final issued diagnosis.
- Outside report text extracted via the Extract Text pipeline.
Tools and repositories
konsultasyon— consultation log.extract-report-text— to parse the outside lab’s narrative for comparison.- Statistical analysis through
ClinicoPathJamoviModule(agreement viameddecide).
How to contribute
- Read the agreement coding rubric (drafted in the repo).
- Pilot-code 20 cases, compare to the reference coder, and flag ambiguous rules.
- Propose rubric refinements with concrete examples.
Pitfalls
- “Disagreement” includes refinement, reclassification, and true discrepancy. Keep the three categories separate in every table; collapsing them hides the meaningful signal.
- Referring-institution identity is sensitive. Analyses are aggregated at the institution-type level (academic / private lab / community hospital), never by name.