Gain isn't always the primary consideration for a PA. It is more common for power added efficiency (PAE) and possibly linearity to be primary design specifications. It is quite possible to achieve high/acceptable PAE without high gain.

That said, if you need a stable high gain power amplifier, then the first question is how power efficient does it need to be. If the answer is "efficiency doesn't matter, but linearity does" then using the Rollet stability factor (k and delta) is an ok approach. But if efficiency matters, then the amplifier is operating in some nonlinear fashion (Class AB, B, C, D, E). In these cases, the K and delta factors change as a function of power level, so you need to measure them with power sweeps using large signal S parameter analysis. Of course the trick is to keep the real part of the total impedance positive over all load conditions; otherwise, the amplifier may oscillate.

I hope other's chime in with their experience and approaches to PA stability.