.jpg){kind=link}
- Published
- Updated
A clinical event known only to the month does not have the precision of a timestamp verified to the minute. Losing that distinction during transfer can create a misleading sequence of events.
Preserve the meaning of data
Define the observed value, time-zone context and any subsequent derivation separately. Automatically inserting the first day of a month is a transformation decision, not evidence that an event occurred on that day. Agree analysis assumptions with the statistics team while preserving uncertainty in the original record. Document where each representation will be used.
Worked example
For example, a hospital records an event as March, but a receiving system requires a complete date. Inserting 1 March could incorrectly suggest that the event began before treatment. In this example, the partial source date is retained, an imputed date is stored separately, and affected chronological comparisons enter a review list.
Review checks
- Check source precision for day, month and time separately during transfer.
- Include daylight-saving transitions and midnight boundaries in test cases.
- Verify that reports distinguish observed dates from derived values.
Acceptance and boundaries
Acceptance means a reviewer can reconstruct both the source information and the assumption added. Conversion to a common time standard may be appropriate, provided the original context remains available. This control does not establish the true onset of a clinical event. It prevents uncertainty from being concealed by processing and keeps corrections traceable with their reasons.
Frequently asked questions
- Must every incomplete date be filled?
- No. The intended use determines whether imputation is needed, and observations must remain separate from analysis derivations.
- What if the time zone is unknown?
- Record that uncertainty explicitly and limit conclusions requiring exact sequencing until additional source information is available.