Yes, frequently. I explained one such case in my edit (which I think made it in prior to your comment but perhaps not).

Well if the treatment depends on the test outcome why wouldn't you just test every case every time?

Because treatment doesn't always depend on the test outcome.

Er so then don’t test in those cases? I don’t see your point.

Yes, exactly. My point? I'm just answering your question, not sure what you're getting at.