the most efficient then you have to take out as many variables as possible. Way, way, way back in the days of film there was a move on for the studios to make the fill the constant. I wish I could remember what the formula was but the idea was a joint effort by the studio and the processor to get the fill to create a specific film density. Once you know the specific f-stop that would create that density then you could work your main light(s) to create what ever effect you desired. It also aided the processor since they now knew that your fill was a specific density and they didn't have to guess. At that time the method of choice to arrive at a fixed fill light was to shoot some hefty strobes into a white ceiling and flood the entire room with an even light. This took lots of watt-seconds but at least you didn't have to be fiddling with the fill lights.
You asked what it the best way so I'm not sure what is the best way. You need to decide what your equipment allows and which do you want to be the variable. Generally I feel the main is the variable and but if you don't have a fixed studio with fixed lighting then they both are variables and must be treated as such.
"Lost"
Part of my DragonScout series
Image created in Poser and PS2