Question 12 of 20 Intermediate Interpretation
What can go wrong when time zones are normalized in dosing and adverse event timelines?
Short answer
Converting local times to one standard can move an event to a different date, reverse the order of dose and event, or shift a pharmacokinetic sample out of its window. The fix is to keep the original local time and the time zone, store a clearly labeled standard time, and derive intervals from one consistent basis.
Why
E6(R3) section 4.2.2 asks that automatically captured dates and times be unambiguous, giving UTC as an example. EMA's computerised systems guideline asks for UTC, or time with time zone, and says users must not be able to change the device clock. FDA's question and answer guidance asks sponsors to state the time zone of timestamps.
Trouble starts when systems mix conventions. An ePRO device records in UTC, the EDC holds site local time typed by staff, and the laboratory system uses its own server time. A participant in New York who doses at 21:30 local time on 3 May and has a reaction at 23:00 has both events on 4 May in UTC.
If the dose is stored as local and the event as UTC, the event looks 5.5 hours after the dose instead of 1.5 hours. Daylight saving changes add an hour of error twice a year. For study day calculations, local date usually matters, because the protocol counts calendar days at the site.
For intervals, both times must be on the same basis.
A strong answer covers
- Keep local time and time zone, plus a labeled standard time
- Date shifts and order reversal across midnight
- Mixed conventions across EDC, ePRO and lab systems
- Daylight saving errors
- Study day from local date, intervals from one basis
Common mistakes
- Converting everything to UTC and discarding the local time.
- Assuming all systems record times in the same zone.
- Calculating study day from UTC dates for participants far from Greenwich.
The interviewer might ask next
How would you check a dataset for time zone inconsistencies before analysis?