As you suggested, there are two solutions: Jean Bélanger's Peak and Hold Board and inline resistors. My professional opinion is that the P&H board is the superior option; the primary benefit being that you offload PWM responsibilities to a chip that makes duty cycle determinations on the fly. Furthermore, all of the noisy injector switching now takes place on an isolated card.

The resistors, on the other hand, reliably work. The problem is that injectors open slower since you're eliminating the current peak. This behavior hurts overall resolution and makes idling larger injectors very difficult.