Yes. As Fred mentions - you process the masters as you normally do.
Clipping masks are great and provides good flexibility. I've used them a few times. Make sure you set the layer folders in pass-thru mode otherwise it doesn't work. I have however had some issues with this process in getting it to display for full colour spectrum or range. Use, plenty of adjustment layers and link them to the layer below can give you better control over this. I would suggest labelling the layers as you go as it can get confusing when you've got 12 or more layers, folders and adjustment layers all doing things to an image.
Another method is to do the convential MaximDL or CCDStack pixel math. This is what I prefer to use, but is perhaps a little more involved (well, probably about the same level as the photoshop clipping masks). Using CCDstack's colour create or MaximDL colour combine tool, start with a weight ratio of 4:1:3 for SII, Ha and OIII. In CCDStack, you have to create the image in order to see it (i.e. there are no previews) and are able to make adjustments to the weights again. MaximDL is different as you can preview the image before applying the weights. In both cases automatically normalise the background colour (this step is no different to balancing an RGB image). Once normalised, arbitarily adjust the R,G,B weights (actually SII,Ha and OIII weights) to get a good spectral range. You may need to weight the SII a little more or drop the Ha to suit your tastes. When the nebula is roughly equal in all three colours and aesthetically pleasing, take note of the R,G,B weights, then cancel the colour combine operation (MaximDL) or close the combined image in CCDStack. For the sake of this post, lets just say the weights were R(SII):4.8; G(Ha):1.0; B(OIII):3.5
Now, you'll need to manually normalise each of the SII, Ha and OIII masters to that the background is equal across them. Using MaximDL or CCDStack, measure the average ADU in the same location across the masters. This location should be void of nebulosity. If a gradient is present, measure the darkest point. Take note of the values between the masters. The goal is to make them equal ADU's around the value of 50. You can use another value, say 100, but the point being is they need to be equal. So, an example; SII ADU is 954 counts, Ha ADU is 1282 counts, OIII ADU is 1183 counts. So to get to 50 ADU, use MaximDL's or CCDStack's pixel math function, you subtract from all pixels in the image - SII:-904;Ha:-1232;OIII:-1133. Once done go back over the SII,Ha,OIII masters and measure the area again - it should now be as close as possible to 50 ADU across the images. They are now normalised are ready for the next step.
Using pixel math again, you want to add to the pixel values across each master based on the weightings you previously noted. So for SII:480(4.8); Ha:100(1.0) and OIII:350(3.5). Once the scaling of the masters is complete using CCDstack's colour create or MaximDL colour combine tool, select the SII, Ha, and OIII masters to be combined, but combine them as 1:1:1 and don't auto normalise the background. You've already configured the weights and normalised the background using pixel math! Voila! You've got yourself a fine looking narrow band image to work from. As Fred mentions, I too like to integrate the Ha data again as Luminance. I've heard of people doing HaSII,HaHa and HaOIII blends, but I didn't see much value in this. A basic Ha,SII:Ha:OIII usually suffices, but probably upsets the narrowband purist. You can of course get a little funky and create a new Ha layer for the luminance alone and stretch it harder or push deconvolution to the maximum.
Should be well on your way...once you get a monochrome CCD and some filters I should add.
Cheers
Nah, I like my OSC too much. No frills imaging and I can do narrowband too. Good all rounder. There's not enough clear skies to do RGB all the time. 
