ORIGINAL: mesae

Some full-scale planes with "flat", metal tube, fabric covered tail feathers are built with constant thickness tube for the stabilizer and leading edge of the elevator, but use thinner tube for rest of the elevator perimeter. This provides some taper, which gives a lighter, more precise control response and feel, and reduces drag (and weight). If those are important to you, then taper. Otherwise, maybe just round the trailing edge. That will help a little to reduce centering force. I don't recommend leaving it square. That will reduce the efficiency of the elevator, increase drag and greatly increase the centering force, which can actually contribute to flutter (excessive centering force can cause overshooting neutral if too strong, setting up a resonance).

For a given established design, the two most important factors preventing flutter are rigidity, and staying below the flutter speed, like Dick said. ALL surfaces will flutter if flown fast enough. We want to build our stuff so the flutter speed is faster than the fastest speed we will ever fly.

Generally:
If it's not rigid enough, it will flutter, regardless of cross-section, if flown fast enough.
If it's rigid enough, it will not flutter, regardless of cross-section.

The cross-section is more important to handling qualities and effeciency than to flutter, but of course there are aerodynamic and mass balancing features that are commonly used to increase the flutter speed for a given system, that also have the benefit of improving control feel/stick forces.

Before someone writes me that this control feel stuff is bunk 'cause it's model, consider again the effect that reduced centering force will have on servo gear and clevis wear and battery life. The forces can be greater than many people realize. More than most people would willingly subject their control surfaces and servos to by applying force by hand, against servo resistance.