The time and cost of a clinical evaluation is very product dependent. Giving two extreme examples:
Bone screws have been in use for many years, there are many international standards and quite a few harmonised ones which you can use to demonstrate compliance with the essential requirements. Consequently there are not many essential requirements that you have to use a clinical evaluation to support your demonstration of compliance. Additionally because they are so well known a lot of the issues have previously been identified. The author of the evaluation should be able to use their experience to highlight the main issues, summarise a few key points from articles and regulatory reports and voila you are done in a few pages, for instance (I'm guessing at these figures) a search of MAUDE will return 1000's of references to bone screws. The author could say that he has reviewed 100 and 80% relate to the head rounding out, 20% to the screw backing out..it can be justified that the sample of 100 is sufficient to identify all but the most obscure of problems
However, if the product is relatively "new" and high risk with less data, lets say a total ankle replacement coated with a HA coating for fixation and TiN for the wearing surfaces then it becomes much more problematic and time consuming because there needs to be more depth to the review...You may only have 500 MAUDE reports on total ankle replacements, but it is a lot more difficult to justify not examining all of them - there are a lot less so it is not so much of a burden, but also, more importantly, because ankle replacements are "newer", it is much more likely that there is an issue the author is unaware of. On top of this the author should also probably examine HA and TiN coatings used in other joint replacements, suddenly what was a few pages for bone screws becomes 10's if not hundreds for a total ankle...both are class IIb!
The other factor that determines the length of the review is the time frame. If this is the first "proper" clinical evaluation then you will have to review all of the relevant data from a suitable time period...maybe when the device type was first introduced?... If an evaluation has already been completed then you only need to review data from when that one was completed-there is no need to review things twice!
One last thing - don't forget to feed the output of the clinical evaluation back into the risk assessment/management!