The thread you linked has a very detailed technical explanation of what’s going on:
What is your ultimate goal? It’s possible to index an entire Scrivener project in DevonThink. That will allow DT to search the contents, but won’t allow DT to add (Scrivener) item-level metadata.
In my own work, I simply use DT and Scrivener side-by-side, with DT as my main research repository. That is, in this case, I might put the interview transcript directly into DT, and then bookmark a DT item link in Scrivener.