[GMB]

There are in-circuit situations where a simple test is possible. The classic example is the coupling capacitor between audio stages in valve circuits, from anode to grid.

If you don't have a high voltage source for the test (without which the results are not reliable) then simply power the circuit and measure the voltage that appears on the grid using a modern high-impedance voltmeter (so not an old AVO). The smallest positive voltage is a very bad sign.

With parallel heaters you can remove the valve as it too can be a source of this kind of leakage current.

Another in-circuit test is to check the whole HT. Ideally you need a circuit diagram as sometimes you find a resistive path from HT to earth. If so you could break it with perhaps only one disconnection and then do a leakage test on the whole HT with valves cold. Be careful to use the lowest working voltage that the HT "sees". This is often a way to bulk check all the screen decouplings.