Other issues with UVC LED-they have to have quartz lens-a metal body so it can be heat sinked-and the efficiency is really LESS than that of mercury and gas discharge sources.And the UVC LEDS are EXPENSIVE-more so than the other sources.
The cost and low efficiency are both linked to the very small size market for these sources today. Today the few applications really benefiting from the LEDs don't mind "paying" the 10x extra input power, when the source is so expensive by itself anyway. But with the prospect of higher volume sales, there will become quite some budget to further develop the construction, plus the higher volume itself will lead to the high fixed cost per batch to be diluted over way larger production batches.
Of course, they will be more complex so more expensive than standard white LEDs, but by far not that ridiculously expensive as they are now.
I don't think they will retain the metal body. That is used today, because it is an universal "of the shelf" package for a low volume, niche optical semiconductor product. I would more guess on ceramic base (same as present white power LEDs), with quartz dome (instead of the plastic blob on the white LEDs), or even a flat piece, cemented onto a spacer ring around the LED chip. Just ideas which were used in the past for similar problems, I'm sure there will be still way better ways to package a strong UV radiator for cheap in mass production...
And second I think their efficiency will be in the 20..40% ballpark, as with the present blue LEDs.
The other question of LEDs for the home disinfection is their safety:
Unlike the discharges, the LED does not inherently radiate the characteristic bluish-grayish light all UV discharges do. So it would be impossible for people to immediately tell if the thing is ON until it is too late. So the device design will have to have extra features to reliably indicate the UV is radiating. It must be a way, which can not fail in the way of not giving the signalization, but still emitting the UV. So a bunch of visible LEDs wont be accepted. Maybe a drop of a phosphor, converting part of the UV to the today characteristic bluish glow. But it should never fade and that is the problem now.
Definitely it can not be anything that becomes obtrusive in normal use (some beeping or flashing or so).
In fact this is a problem already discussed with the present industrial use, where these sources are encapsulated. The danger is seen in the machine failure or maintenance/repair, when an accidental irradiation of a worker may happen. Because it is completely not visible, makes it so dangerous.