I have seen duff patches where the wrong layer peeled off in preference, but this is rare...
I reckon there are two main differences between the bond at the edge of the patch and the bond in the centre
a) the bond in the centre potentially sees higher strains because the patch is stiffer there (whereas the feathered edge is stretchier) and
b) the potential for any residual solvent from the rubber solution to diffuse away from the patch centre is less than at the edge; the edge is made of thin, not very cross-linked rubber, through which diffusion is more likely to occur.
The patch anyway sees more stress if the tube is undersize vs the cover, or the patch is near the tyre bead (which strains the tube horribly).
So it is possible that the rubber solution wasn't 100% dry when you put the patch on and this has resulted in a weakened bond in the patch centre more than at the edges. Modern rubber solution has a higher boiling point solvent in it than the TriCh that was used years ago. Particularly when it is cool and/or humid, it can take a long while for the modern solvent to flash off when applying the rubber solution.
To hasten this process I usually (at home) / often (at the roadside) use a naked flame to drive the solvent off. If you let the solvent burn for more than about one second there is a real danger that the inner tube will have a (much bigger) new hole in it.......
; you need to be ready to blow the solvent flame out almost immediately, and repeat (after a few seconds to allow cooling) until the rubber solution won't catch light any more. Just then, the rubber solution is as dry, as moisture free, and as warm as it is likely to ever be before you apply a patch to it, which significantly increases the chances of getting a good bond.
You will know if you have slightly overdone it because the rubber solution will boil and blister if it is too thick, too wet, or is heated too much.
cheers