Short answer: Yep.
I have a friend who used to work for a big company that did ebook conversion. As little as a year ago, as crazy as this sounds, the workflow included scanning the print-ready .pdf file via OCR. Yes, even if the raw, unadulterated digital text existed.
The rationale, as I understand it, is that their workflow also had a bunch of software the mostly replicated the work of the print book’s interior designer, so that the ebook looked a lot like the print book.
This, he said, was easier than starting with clean digital text and trying to edit the CSS by hand to replicate the print book.
To which I replied — why the fuck are you trying to replicate the print book? It’s an ebook. Stop it.