Seems to me that you would like to know the lineup power amplitudes at each stage, for this. Could be one bad, could be many "just a bit weak" all stacked up. Pout could be depressed by any. How's your small signal RF probing?
Might also look at the problem along calendar lines, like does "yield cliff" appear just after some BOM change or some "non critical" vendor / part (mfr, grade, etc) change trailing by your inventory clearing
cycle.
A hot air desoldering tool or just a soldering iron tip, might give you more localized sensitivity info. But keep the possibility of gang related activity in mind.