I think I get what you are on about, but I dont think that it is possible to develop such a circuit...
The problem is that you are trying to compare the current being drawn by the device, with the 20mA, before the device is activated. So you will always be comparing the 20mA current with zero current because the device isnt being allowed to draw power.
It makes me wonder though, if what I am assuming is true, is that you want to test the wires with a quick jolt of electricity (no idea how the testing is working), if it would be possible to use a relay in parallel with the switch, near the appliance, that would only allow the device to turn on when there is a continous circuit attached? No idea if this would work though.
Although I do have to say, I cannot see a reason why this sort of circuit would need to be used. I was under the impression that in testing the cables, you would have to follow some form of strict guidelines to do it properly?