thebench.bio #1 Ecosystem for the Bio-Tech Industry
All Quality Assurance questions

Question 10 of 20 Fundamentals Practical

How would you review an executed batch record before release?

Short answer

Check that every step was done as written, recorded at the time by the right people, with values inside limits, yields within range, and every deviation or discrepancy investigated and closed or assessed. Then confirm testing, labeling and reconciliation are complete before recommending release.

Why

US regulation requires the quality unit to review and approve all production and control records, including packaging and labeling, before a batch is released, and any unexplained discrepancy or yield outside limits must be thoroughly investigated. Review is not proofreading. It asks whether the record proves the batch was made under control.

A good reviewer reads the record against the master instructions, looks at the original data behind key entries, and asks what is missing as well as what is wrong. Entries recorded at the same minute for steps that take an hour, identical handwriting for two different operators, or a printout that does not match the written value all deserve questions. Yields matter because a low yield can mean lost material and a high yield can mean an extra addition or a counting error.

The review should also connect events: an alarm in the equipment log should match a deviation in the record. The aim is to find problems before the QP or quality unit decides, not to annotate them afterwards. Review by exception, where an electronic system flags only out of range entries, is efficient but only as good as the limits and alarms configured in it.

How to work it through

  1. Confirm you are reviewing the current approved master record version for this batch.
  2. Read each step for completeness, signatures, dates and contemporaneous entries.
  3. Compare critical values and printouts with limits and original data, including audit trails where electronic.
  4. Check yields and reconciliations against their ranges.
  5. Confirm every deviation, alarm and discrepancy is linked to an investigation and assessed for this batch.
  6. Verify test results, labeling records and line clearances are complete before signing the review.

An example

A reviewer opens a tablet batch record of 64 pages. Page 22 shows granulation end point at 14:05, but the mixer printout shows 14:35 and an amperage alarm at 14:20. There is no deviation for the alarm.

The compression yield is 96.8 percent against a range of 95 to 101 percent, so it passes. The reviewer raises a deviation for the unrecorded alarm and time mismatch, holds the batch, and asks production to explain before the record goes to final approval.

Where the answer stops holding

Record review finds what was recorded, not what was done off the record, so it relies on a culture of honest entries and on audit trails. Review by exception misses problems that sit inside configured limits. Very long records invite fatigue, so risk-based focus on critical steps helps but must not become skipping.

A strong answer covers

  • Compare against the approved master record
  • Contemporaneous, attributable entries and original data
  • Yields and reconciliations within limits
  • Every discrepancy linked to an investigation
  • Testing and labeling complete before release

Common mistakes

  • Checking only that every signature box is filled.
  • Ignoring equipment alarms or printouts that are not mentioned in the written record.
  • Accepting a yield inside range without asking whether it is consistent with the history.