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

Question 13 of 20 Intermediate Troubleshooting

How would you distinguish an isolated documentation error from a systemic quality-system weakness?

Short answer

Look beyond the one record. Search for the same error type across people, shifts, products and time, and examine the form, procedure, workload and review step that let it through. An error that recurs or that review routinely misses points to a system, not a person.

Why

A single missed signature can be a slip by one person on one day. It can also be the visible tip of a form that invites the error, a procedure nobody can follow at the line, or a review step that no longer looks. The way to tell them apart is to widen the sample before deciding.

Search the deviation system and recent batch records for the same field left blank or the same correction pattern. If it appears across several operators, the person is not the cause. Check whether the reviewer caught it, since an error reaching QA after two reviews tells you the review is not working.

Look at timing, because errors that cluster at shift change, at month end or after a staffing cut point to workload. ICH Q10 expects deviations, CAPA and monitoring to feed one another, so a classification of "isolated" should rest on evidence that you looked for recurrence. The MHRA notes that culture and performance targets shape what gets recorded, so a site where people fear blame will produce tidy records and hidden errors.

If the search finds nothing similar in a reasonable sample and the cause is understood, the isolated label is fair. If you cannot search because records are paper and scattered, say so rather than assuming the best.

How to work it through

  1. Define the error precisely, such as an unsigned verification at step 14.
  2. Search a defined period of records and deviations for the same error and close variants.
  3. Break findings down by person, shift, product, form version and reviewer.
  4. Examine the form, procedure and review step that should have prevented or caught it.
  5. Decide isolated or systemic on that evidence and document the search.
  6. Choose a correction for the record and, if systemic, a CAPA aimed at the design rather than the individual.

An example

QA finds an unsigned second-person check for a component addition. A search of 90 batch records over three months finds 7 more, spread across 5 operators and all on form version 4, where the verifier box sits on the next page. None was caught by production review.

The pattern rules out one careless person. The team moves the box beside the step, adds an electronic hold until verification is entered, and tracks the error rate, which falls to zero in the next 120 records.

Where the answer stops holding

A search only finds errors that were recorded or noticed, so a clean result is weaker evidence than it looks. Small samples can miss a low-frequency systemic problem. Electronic records make the search easy, while paper records may limit how far back you can practically look.

A strong answer covers

  • Search for recurrence across people, shifts and time
  • Check whether review caught the error
  • Look at form, procedure and workload design
  • Document the evidence for the classification
  • CAPA on the system, not retraining one person

Common mistakes

  • Classifying an error as isolated without searching for others.
  • Concluding the operator is the cause when several operators made the same error.
  • Ignoring that two reviews missed the error before QA found it.