[Placidus (Mike and Trish)]

That seems a perfectly legitimate approach, to me. I like to think that in amateur astrophotography most of us are interested in at least three things: showing the structure of the object, showing something about the astrophysics of the object, and producing a beautiful, artistic image. I like to think that we can do all three at once at least to some extent, without being too precious and legalistic about it. To that end, you've succeeded completely.

You are arguing that (1) the H-alpha emission line really is red, so it is more information about the red emission from the object, and is legitimate to add to the red channel. (2) To the extent that H-alpha is ubiquitous, you can use H-alpha as a kind of super-sharp star-reduced seeing-robust luminance signal. (3) Having made that decision, applying the H-alpha luminance signal to your blue channel will tell us where there is OIII but also H-alpha - the intersection of the two.

The first step, arguing that in an emission nebula, the red is mostly due to H-alpha, and the red filter and the H-alpha filter are kind of interchangeable, is obviously not really correct in the sense that you'll get a different result from if you just used tight narrowband filters (and red continuum, NII, SII, and some others are gone), and you'll get a different result from if you just used a normal red filter (which would include the lot). But if you just argue that you're after an RGB image of the Lagoon where you've emphasised the H-alpha, nobody can disagree. I think it works every well with some galaxies, like say NGC 300, if you want an image of the galaxy with the emission regions enhanced or exaggerated.

That third step won't work so well in some other contexts, for example parts of the Magellanic Clouds, where you can have regions of very high vacuum, very high energy, where there is lots of OIII emission but very little H-alpha, and you would not capture those regions even though they are there and important.

In summary, I think it's completely legitimate, to the extent that you and the viewer are aware of the implications.

I have some difficulties with images where people make quite arbitrary mixes of different channels - a kind of polypharmacy - and it's not clear what the resulting image actually means astrophysically. It can be good art, but there are ways of producing good art without losing touch with meaning.