QMS for Life Sciences
Elsmar Cove Forum Sponsor

Design Verification & Process Validation - Statistical sample sizes

Nancylove

Starting to get Involved
#1
Hi all:

I’ve seen a lot of threads of Design Verification, Process Validation and statistical sample sizes, but I haven’t seen anything that answers my question. I’m looking forward to input from experts in these areas regarding to my questions below.

If you have a feature such as 2 plastic tubes that are bonded together and a product requirement of tensile strength (ex: 5 N) for the bonded assembly, should you do both design verification and process validation for the same requirement? Design verification is required per 7.3.6 of ISO 13485, and since it’s a mfg process, Process Validation per 7.5.6 is required also. What is the purpose of doing Design Verification for this feature if Process Validation will provide the final results? Why not perform Process Validation to satisfy the 7.3.6 requirements?

Note: I did find one post from 2008 in which the conclusion seems to support Process Validation can satisfy Design Verification, but I'd like the latest input.

Thank you.
 

Ronen E

Problem Solver
Staff member
Moderator
#2
You might receive (useful?) input here but the only expertise you will find here is a self-professed one. It is solely up to you to determine the real worth of any advice you find here.

On point:

Design Verification (DV henceforth) and Process Validation (PV) are intended to serve different purposes and occur at different stages of the D&D endeavour. After the case, a proper PV might technically satisfy the requirements of DV, but in the majority of cases its (proper) timing will completely miss the purpose of DV. Typically, DV occurs at the prototype stage, when the commercial production process doesn't yet exist or is just starting to form. It usually involves a relatively small unit count (typically prototypes), is not intended to provide full assurance, and is targeted at gaining sufficient confidence that the design, as at the time arrived at, is satisfactory (meets the latest approved design input). The manufacturing techniques employed may be provisional and quite different from the scaled up production techniques that will be in place for PV (typically, quite some time later). A fully successful DV may trigger design freeze, setting up a pilot- or an initial full-scale production line, progressing to design validation (sometimes blended with PV) etc. But the DV might also end up being less than fully successful, and trigger design input updates or design changes (which will typically lead to another DV iteration).

PV doesn't have to take place at all. It's required only where 100% production verification can't/doesn't happen in a specific process. Where it does take place, the purpose is to prove that the established production process is reliable (actually, part of the process establishment may occur as part of PV). For PV, the design is already not in question.

Perhaps you will find interest in my article about design verification.
 

Nancylove

Starting to get Involved
#3
You might receive (useful?) input here but the only expertise you will find here is a self-professed one. It is solely up to you to determine the real worth of any advice you find here.

On point:

Design Verification (DV henceforth) and Process Validation (PV) are intended to serve different purposes and occur at different stages of the D&D endeavour. After the case, a proper PV might technically satisfy the requirements of DV, but in the majority of cases its (proper) timing will completely miss the purpose of DV. Typically, DV occurs at the prototype stage, when the commercial production process doesn't yet exist or is just starting to form. It usually involves a relatively small unit count (typically prototypes), is not intended to provide full assurance, and is targeted at gaining sufficient confidence that the design, as at the time arrived at, is satisfactory (meets the latest approved design input). The manufacturing techniques employed may be provisional and quite different from the scaled up production techniques that will be in place for PV (typically, quite some time later). A fully successful DV may trigger design freeze, setting up a pilot- or an initial full-scale production line, progressing to design validation (sometimes blended with PV) etc. But the DV might also end up being less than fully successful, and trigger design input updates or design changes (which will typically lead to another DV iteration).

PV doesn't have to take place at all. It's required only where 100% production verification can't/doesn't happen in a specific process. Where it does take place, the purpose is to prove that the established production process is reliable (actually, part of the process establishment may occur as part of PV). For PV, the design is already not in question.

Perhaps you will find interest in my article about design verification.
Thank you. Your comments makes sense.
 
Top Bottom