Three identical numbers can have different histories
A weight could be measured in a clinic, reported by a patient, or copied from an earlier visit. The displayed value may be identical in all three cases, while the evidence about current body weight differs.
Without provenance, an analysis may treat each copy as a new measurement.
Useful details to preserve
- The originating source or device
- The event and recording times
- Whether the value was reported, measured, calculated, or imported
- Transformations such as unit conversion
- Corrections and the version used for analysis
FHIR includes a Provenance resource for recording relevant creation and change context. Using that resource does not guarantee the assertions are correct; provenance records also need trustworthy processes.
Follow a surprising value backward
If a result looks implausible, tracing the source can reveal a unit conversion, duplicated import, or later correction. Replacing it silently with a plausible number destroys that trail.
For a report, identify the data version and the transformation rules. That makes it possible to explain why a later analysis differs. Provenance is especially useful when a polished dashboard hides several stages between the original observation and the number a reader sees.
