Replacing a defective part is no big deal, just incorporate an inspection of the part into a regular maintenance window every X cycles (a 'cycle' is usually one flight) and model its cost as an amortized per-cycle cost based on its mean time to failure.
The problem is when it's difficult to know whether a part is defective. Something like subsurface delamination might cause visible bubbling or warping, but if it's deep enough then it may only be apparent on an ultrasound or X-ray scan. That sort of scan might be more expensive than just replacing the part regularly, and shipping it off to a factory to be inspected and refurbished.
That is a maintenance technician's worst nightmare: an expensive and bulky part that fails in invisible ways, requiring either regular replacement or time-consuming inspection with expensive equipment.
