Getting from the Earth's surface to LEO will be an expensive proposition for a very long time. That's an unavoidable fact of life. But the cost of launching for any large economy is still a tiny fraction of overall GNP. By putting up as much infrastructure into space as possible, the cost of travel once in space will drop considerably.

Solar and nuclear thermal rocket engines could be built for in space applications. These engines would not require cryogenic propellants and no oxidizers. Any type of gas could be used as reaction mass, simply heat and expel. This would make them extremely simple, reliable, and flexible in a wide range of environments. The first step may be expensive and a hard one to take, but it's well worth it IMO.