Can a reviewer ask for the primary data (actual collected data in a prevalence study)? Is the information confidential or can be shared? Even if can be shared, does the reviewer have the authority to ask for it?
It's a very free process. The reviewer can ask for it if they want (though they would have to keep it confidential). You can refuse if you want. The reviewer can, as a result, decide to give a poor review. The editor can decide to reject the paper on that basis.
In a few cases there might be a good reason that you would not want to share your primary data, but generally, I don't see why it would be a problem. In any case, many journals (at least in my field) require, or at least strongly recommend, primary data to be made publicly available upon publication.
If you are worried about being scooped, you are probably worrying over nothing. If a reviewer was to try to present your data as their own (very unlikely and difficult to start with), it would be very easy to demonstrate, through your documented submission, that the work was yours and they were stealing it.