>Ok. Here's my problem. We have a field out there called CDC-Id. It's the only identifier that we CAN and DO have for a patient. The CDCID is system generated on the center's end (which are all over the nation). The submit the forms to us and we input them into the system. Now two things can happen....as you very well can already guess, 1. the nurse/user on the center's end can enter in the wrong id and/or 2. the data entry person who receives the form can enter in the wrong id.
>
>When this happens, we must go into the system and make the changes to the cdc id. Now if we make this the primary key, doesn't this mean that we won't be able to update it?
>

You can make your PK updatable, there is nothing technically wrong with that. But the inherent problem in your design is the user's ability to edit the primary key. Its considered good design practice to give each table a surrogate primary key- that is a key with no real world meaning, so the user will never have a reason to change it (the user never sees it). You can still enforce uniqueness on your CDCId, it would just become a candidate key instead.