One possible method could be finding suitable Metal-Oxide (M-O) vibrational mode, which is conveniently active in Raman spectra, where position of spectral line depends on the oxygen content in your samples. As oxygen vacancies in your samples do influence crystal structure, its periodicity and therefore vibrational modes of crystal matrix, correlating this with other spectroscopic and structural characterization methods could give you the answer you are looking for.

I'm not sure about metal-oxides but very similar thing has been done in the following article where the position of Si-O vibrational stretching mode has been found strongly correlated with oxygen content, and verified with other characterization methods.

I hope this will help you out.

Article Local Site Distribution of Oxygen in Silicon-Rich Oxide Thin...