Search the Elsmar Cove!
**Search ALL of** with DuckDuckGo including content not in the forum - Search results with No ads.

"Process Verification" of manual process


Starting to get Involved
So we have a business unit that makes very small batch (typically 1-2 pieces) custom orders, and the process is primarily manual assembly. We were going to add what is essentially a new feature offering to the product line (note that this process is already previously in place at a sister site- we are following the same work instructions, but we are now also manufacturing it).

There was an engineer assigned to implement this new product offering, and he is no longer with the company, but he gave me the meat and potatoes of his "qualification" activities, which he essentially called a "process VERIFICATION report". In it, he essentially has the new product offering styles built and then inspected per the routine production inspection requirements established at our sister site. Note that we would be doing 100% manual inspection to the same inspection criteria when we go to full production.

I know that if this were an automated process we would typically do process VALIDATION. But since it's not, and we will be doing future 100% inspection of production products anyways, is this "process VERIFICATION report" necessary? Should something else be done instead?

This is med device industry if it makes a difference.

John Predmore

Involved In Discussions
There is a subtle dividing line between verification and validation. Some companies call these overlapping activities ver/val, V&V, or use the generic term “qualification”.

The decision of verification versus validation is not whether the operation is manual or automatic, but whether quality output can be readily determined. One example where validation may be required is when testing is destructive, so the effectiveness of the operation must be demonstrated with statistical confidence in controlled experiments. Another example is when alignment or interaction of a feature installed in this station can only be ascertained on the finished assembly, so again, careful planning and experimentation are needed to validate.

You mentioned a new feature added to the process. It is possible to do an incremental ver/val (sometimes called a “delta qual”) where only those portions of the operation affected by the change go through ver/val. The alternative would be to do a complete ver/val of this operation. Be alert to the possibility of unintended consequences where something you innocently add in station #1 causes interference at one or more downstream operations.

Process verification and/or validation should be examining more than product output. Part of ver/val is that all revisions to work instructions and other documents are complete, in-place, adequate for their purpose, and followed by the workforce. If new tools, labels, bins or fixtures were needed to add the new step to the process, are those adjuncts in-place and adequate? If setup, calibration or maintenance of equipment or tools are affected by the additional step, are revisions in-place and adequate?

You ask whether something else [or additional] should be done. That question should be decided by your validation team. I recommend there always be more than one person involved in ver/val decisions. There is an inherent conflict of interest when the engineer who validates is under pressure to get the manufacturing process up and running quickly. The countermeasure against subconscious bias is a cross-functional team reviewing and signing off on all ver/val plans, execution, and reports.

Top Bottom