Topic has been retitled for clarity and archival purposes. Please make ongoing responses to messages with the new title.
I actually did not know that NVDA OCR presents a page at a time. This is most likely because my standard advice is to do OCR processing before opening a document, usually PDF, in a screen reader. Using that technique creates a text layer for the entire document prior to it ever being handed to the screen reader, and you get a result that's the one you're asking for here.
Brian - Windows 10 Pro, 64-Bit, Version 2004, Build 19041
If you think that you can think about a thing, inextricably attached to something else, without thinking of the thing it is attached to, then you have a legal mind.
~ Thomas Reed Powell