Centaurus... Do you have any idea how I might find out more about that B707 prang at Manila?
Tinny... No mate, I've been deliberately using the terms "invalid" and "erroneous" to differentiate between this problem and the more widely known "false" indication. A "false GP" is one caused, as you quite correctly state, by side lobe signals. Theoretically if you intercept it and could follow it, it would still produce the normal GP angle of descent.
An "invalid" or "erroneous" GP, OTOH, is quite different in that it is only present during maintenance work, or during certain calibration activities. The radiated signal is deliberately skewed, to test various signal parameters and this is done by navaid technicians. They are supposed to reset the equipment after those activities.
Sometimes they forget to do so. Even then, it only becomes a significant problem if the relevant monitors in the ATC facility are U/S, because then ATC depends on the techs to tell them when they're mucking around with the signal. Sometimes, this communication doesn't occur.
The event at Apia had all these things conspiring at the time and, worse still, the navaid tech's own monitor had been left in "bypass" mode so that it couldn't automatically shut down the defective GP Tx and transfer control to the standby Tx.
You don't counter a spurious side lobe by means of altitude and distance checks, because (all things being equal) it will still produce a 3-degree (or so) descent path. You counter it by ensuring that you're inside the capture zone of the true GP signal. Instrument approach procedures generally account for this n one way or another and ATC radar vectors are another line of viable defence.
BOAC... I would hope that things like that are detected during the ILS commissioning, or periodic inspection, flight checks. Procedures can then be developed to get around the problem, one way or another, and ensure that pilots are aware of it.