01Identify the departure in concrete terms

A useful statement compares what the procedure required with what actually occurred. It names the affected activity, sample or run, and the point at which the departure was recognised.

The third example is more specific than the first two, but it still does not decide validity. The impact depends on the particular preparation, procedure and supporting evidence.

Keep a process departure distinct from a specification exception. A test can follow its procedure and give an OOS result, or depart from its procedure while producing an apparently passing value.

02Ask how the event could affect the measurement

In an original illustration, an analytical solution is held longer than the method’s documented period before measurement. The immediate question is whether that additional interval could change the measured property under the actual conditions.

A statement that the final number passed does not answer that question. A biased measurement can fall inside a broad acceptance interval. The assessment needs evidence relevant to the stability or suitability of that preparation.

Alternatively, a departure may concern an administrative step that did not alter material handling or data processing. Its impact should still be explained rather than inferred from the category assigned to the event.

03Separate correction from prevention of recurrence

ICH Q10 describes structured investigation of deviations and other quality signals to identify causes and inform corrective and preventive action. This pharmaceutical quality-system model also calls for evaluating the effectiveness of those actions.ICH Q10 — Pharmaceutical Quality System (opens in a new tab)

Correcting a record, repeating affected work and changing a procedure are different actions. One may address the immediate output; another may address why the event happened. A statement that staff were reminded of a procedure does not show which cause was established.

For the hold-period example, a new preparation might supply a usable replacement result. It would not by itself explain why the earlier preparation exceeded its period or establish whether other runs were affected.

Read the action against the cause. If the cause was unclear scheduling, a remedy should have a plausible connection to scheduling. If no cause was established, the report should avoid presenting a guessed cause as confirmed.

04Read what closure actually establishes

A closed deviation record means the organisation has completed its stated handling of the event. To understand the analytical consequence, look for the decision about affected results, any replacement work and the remaining limitations.

If an issued result changed, identify the revised report and its connection to the original. If the result was retained, read the evidence supporting its continued use. Closure and result validity should not be collapsed into the same assertion.

A useful reader summary names the departure, affected scope, supported impact and action taken. It can acknowledge that an investigation was completed while still identifying a missing explanation in the documents available to the reader.