jmullins: I don’t exactly understand your problem. You need to convert the original docx to use {unformatted} citations [Tools > Convert to Unormatted Citations] before import to Scrivener. Then use {unformatted} citations in Scrivener, then compile in scrivener and scan in Endnote. I only use Endnote X8 when forced to by a collaborator (much preferring Bookends), but I’ve never had any problems with anything missing or with scanning for final output (I prefer to do that in Word 2016). You most likely need to “re-lookup” refs is because you have different libraries open between insert and scan (thus the #uniqueid is different).
You have probably been using Endnote’s CiteWhileYouWrite. You have to use the Endnote function in Word to “unformat” your paper – to return its citations to Endnote’s non-WYSYWIG form (“temporary cites”) – these are plain text and look like: {Burgess, 1987 #1728} – and Endnote’s CWYW hides them in code that will not come across. Then import your doc into Scrivener.
Because temporary cites have the Endnote record number in them, later when you process your output from Scrivener (either directly as rtf or as docx in Word), Endnote does not need to ask you to disambiguate any of your citations.
-gr