We make quite sophisticated Word documents, containing a lot of bookmarks, cross references, automatic counters, indexes, table of contents, literature references, etc.
For music we used Dorico snippets in png format, but we realised that using SVG exports gives much better results (because it’s vector graphics of course). Drawback: even when you save the Word document in PDF format, you are not 100% sure that your SVG renderings will be displayed correctly on another OS platform (example: you generate the PDF on Windows, and you look at the PDF on Mac: not sure the result will be exactly the same; see this thread for example).
@ZeroZero , I think your Word solution is the correct way to go. And high resolution graphics snippets are probably the most practical workpath.
If it is really very demanding, maybe look at @dan_kreider 's solution for his magnificent hymnal that he created with Indesign. But maybe there are Word features that you absolutely need, and that are missing in Indesign.