So efficiency is always going to be some A divided by some B. Traditionall with rocket engines it means impulse divided by propellant mass expended. But given that there are both electric thrusters with much higher ISP and electric thrusters with much lower ISP (e.g. resistojets) I'm not seeing any reasonable efficiency metric under which I think the claim is plausible. Even in terms of losses electric thrusters that don't have to ionize their propellant have a pretty big advantage.