Every kind of pump will work best with a predictable, constant discharge pressure. What I've found is that at 3,000 ft if I'm pumping water, the discharge pressure should be around 1,300 psig. If I have near zero psig on the tubing, it is 1,300 psig as long as there is no free gas in the discharge stream. Add as little as 1 MCF/d and (with zero psig surface pressure) the pump discharge drops to under 50 psig because of the space required for 1 MCF/d. If I add 200 psig back pressure then the pump discharge goes to some value above 1,200 psig.
Pumping light oil is even worse because of the certainty of phase change at low pressures.
With a sucker-rod pump, if it ever gas locks then leakage past the plunger is required to break the gas lock. With 1,200 psig pump discharge you'll break the gas lock in something like 6 hours. With 50 psig it will take closer to 6 days.
With your drop in effeciency, I'd say that the standing valve is only opening 1/2 to 2/3 of the strokes. Yeah, I'd recommend a back-pressure valve.
David