How would you make a batch disposition decision when the investigation is not finished?

Short answer

Keep the batch in quarantine until the investigation supports a decision, and do not let supply pressure substitute for evidence. If a decision must be made earlier, define what evidence is needed, state the residual uncertainty, and record a risk-based decision that QA owns and that can be reversed.

Why

US regulation requires production records to be reviewed and any unexplained discrepancy thoroughly investigated before release, and the investigation must extend to other batches that may be affected. EU Annex 16 lets a Qualified Person certify a batch with an unexpected deviation only if registered specifications are met, the deviation is thoroughly investigated, the root cause is corrected and a risk assessment concludes the impact is negligible. Investigations pertaining to the batch must be complete to a level sufficient to support certification.

That phrase is the key to this question. Not every open action blocks release, but every question that bears on this batch's quality must be answered. ICH Q9(R1) adds two useful ideas.

Risk to quality includes loss of product availability that harms patients, and the rigor of a decision should rise with uncertainty and importance. So a shortage of a critical medicine is a legitimate factor in how fast you work, never a reason to accept unexplained risk. In practice, separate questions about this batch from systemic ones.

If the open item is why a mixer stalled, but data show the blend was uniform and all attributes pass with margin, the batch question may be answered while the systemic CAPA continues. If the open item is whether sterility was compromised, it is not. Escalate to management when supply consequences are serious, and document who decided, on what evidence.

A strong answer covers

Quarantine is the default while evidence is incomplete
Separate batch-specific questions from systemic follow-up
Annex 16 conditions for certifying with a deviation
Rigor scales with uncertainty and importance under Q9(R1)
Documented, QA-owned decision with escalation

Common mistakes

Releasing because test results pass while the deviation that could affect them is unexplained.
Treating supply pressure as justification rather than as a reason to resource the investigation.
Refusing to release until every CAPA action is closed, even when none bears on the batch.

The interviewer might ask next

What would you write in the release record so a later reviewer can understand why you decided as you did?