thebench.bio #1 Ecosystem for the Bio-Tech Industry
All Process Validation questions

Question 12 of 20 Intermediate Troubleshooting

A validated process drifts gradually but stays within specifications. How would you respond?

Short answer

Treat it as a Stage 3 signal that the process may be leaving its validated state. Confirm the drift statistically, find what changed, assess capability and the time left before failure, and act through CAPA or change control before a batch fails rather than after.

Why

The FDA's 2011 guidance says the goal of continued process verification is continual assurance that the process remains in a state of control, using statistical trending that guards against overreacting to single events but also against failing to detect unintended variability. A validated state is not just passing specifications. It is the process behaving as it did when validated.

A slow drift, such as dissolution falling 1 percent per quarter, means something has changed even if every batch passes. Start by confirming the drift is real with control chart rules rather than eyeballing, and check the measurement system, since a drifting method can look like a drifting process. Then look for what changed around the start: raw material lots or suppliers, equipment wear, a new operator group, a seasonal humidity effect or a minor change that was judged not to need validation.

Recalculate capability on recent data. A Cpk falling from 1.8 to 1.1 tells you the margin is shrinking. Estimate when the trend would cross a limit, because that sets the urgency.

Annex 15 asks ongoing process verification to consider incremental changes over time and whether actions such as enhanced sampling are needed. The response might be a CAPA to correct the cause, a change to restore the process, or increased monitoring while you investigate.

How to work it through

  1. Confirm the drift with predefined statistical rules on the CPV chart.
  2. Check the analytical method for a parallel drift before blaming the process.
  3. Build a timeline of changes to materials, equipment, people and environment around the start of the drift.
  4. Recalculate capability on recent batches and project when a limit would be reached.
  5. Increase sampling if the margin is small.
  6. Correct the cause through CAPA or change control and confirm the trend reverses.

An example

Monthly CPV review shows tablet hardness falling from a mean of 120 N to 104 N over 18 batches, against a lower limit of 90 N. Dissolution is still compliant but trending up. The method is unchanged and control samples are stable.

The timeline shows a new lot sequence from the microcrystalline cellulose supplier began at batch 1 of the drift, with particle size near the upper end of its specification. Capability for hardness has fallen to Cpk 1.0. The team tightens the incoming particle size range through change control and adds compression force adjustment within the validated range, and hardness returns to 118 N within six batches.

Where the answer stops holding

Drift can be caused by several small changes at once, making root cause hard to prove. Trend detection needs enough batches, so infrequent products may drift for a long time before it is visible. Restoring a parameter within its range is acceptable, but chasing the drift with repeated adjustments can hide the cause.

A strong answer covers

  • Validated state means consistent behavior, not just passing
  • Statistical confirmation of drift
  • Rule out measurement drift
  • Timeline of changes and capability trend
  • Act through CAPA or change control before failure

Common mistakes

  • Ignoring the trend because every batch still passes.
  • Assuming the process drifted without checking the analytical method.
  • Repeatedly adjusting settings to hold results without finding the cause.