How to Handle a Protocol Deviation
Classify, document, and act on a deviation before it becomes a pattern or a serious breach.
What a how-to is not
A how-to is SPEQ’s practitioner method, not a procedure. It does not replace your own SOP, it is not a validated approach, and the judgement calls in it belong to your quality unit.
Deviations are not failures of discipline; they are the normal friction between a written protocol and real participants. What distinguishes a well-run trial is not the absence of deviations but whether they are found quickly, classified honestly, and acted on when they cluster. The finding that hurts is not the deviation — it is the log that shows the same deviation forty times with no response.
- 1
Record the deviation when it is found, not when it is resolved
Capture what happened, when, which participant and visit, who identified it, and how. A log entry created after the corrective action is complete cannot show how long the condition persisted, which is the fact that determines its significance.
- 2
Classify by impact, not by inconvenience
The question is whether the deviation affected participant safety, participant rights, or the reliability of the data. A missed non-critical assessment and a dose given outside the protocol window are not the same category, and grading everything the same defeats the purpose of grading.
- 3
Escalate the ones that meet the serious-breach threshold
A breach likely to affect the safety or rights of participants, or the reliability and robustness of the data, has a reporting obligation with a short clock and is escalated on discovery rather than at the next review. Deciding this at a monthly meeting has already missed the point.
- 4
Assess whether the participant may continue
Some deviations bear on eligibility or on the safety of continued participation, and the investigator must decide, with the sponsor, whether the participant continues. Enrolling a participant who did not meet eligibility is not corrected by noting it — the continuation decision is its own documented judgement.
- 5
Look for the pattern before writing the corrective action
One deviation is an event; the same deviation across several participants or several sites is a protocol or training defect. Corrective action aimed at the individual occurrence when the cause is an unworkable visit window will produce the same deviation next month.
- 6
Feed the analysis and the study report
Important deviations are summarised in the clinical study report and may inform sensitivity analyses. A deviation log maintained separately from the reporting and analysis path gets reconciled at database lock, when it is too late to do anything about what it shows.
- !Deviations logged at resolution rather than discovery, so the duration of the condition is unrecoverable.
- !Uniform classification, which hides the few deviations that actually bear on safety or data reliability.
- !A serious breach held for the next scheduled review instead of escalated on discovery.
- !Corrective action written against a single occurrence when the log already shows a systemic cause.
How to Handle a Protocol Deviation: frequently asked questions
Common questions on handle a protocol deviation.
What makes a deviation important rather than routine?
Whether it could affect participant safety or rights, or the reliability and robustness of the trial results. That test is applied to the specific occurrence, so the same nominal deviation can be routine in one participant and important in another — which is why blanket classification by deviation type is unsound.
What is a serious breach?
A breach likely to affect to a significant degree the safety or rights of participants, or the reliability and robustness of the data generated. It carries a reporting obligation on a short clock from the moment it is discovered, and the assessment is made when it is found rather than deferred to a periodic review.
Should a participant who deviated be withdrawn?
Not automatically. The investigator and sponsor decide whether continuation is safe and whether the participant still meets the conditions the protocol requires, and they record that reasoning. Withdrawal is one possible outcome, not the default remedy, and neither continuing nor withdrawing is defensible without the documented judgement.