The power dissipated in Q1 depends as much on the voltage across it as the current through it. There is no linear regulator which can defy this fundamental law of physics.
Q1 starts to conduct when the load current is about 200mA, you could reduce R1 so more current goes through the regulator before it starts to conduct but I'm guessing your real problem is lack of heat sink. The heat dissipated by the regulator and transistor combined is (Input voltage - Output voltage) x Input current. You can't change that I'm afraid, the best you can do is get rid of the heat with a heat sink.
Brian.