The polar curve r = 2\theta is a spiral in the counterclockwise direction. If we're talking about the slope of the tangent line at (pi, pi/2) (polar coordinates), we have to be talking about dy/dx, because the other derivative, dr/d\theta is constant and equal to 2. I drew a quick sketch of this curve and convince myself that dy/dx at the point in question was negative, which agrees with the answer you gave, at least in sign.
You need to convert your polar equation into Cartesian coordinates, and then take the derivative dy/dx.
Since \theta = tan-1(y/x), the polar equation becomes
\sqrt{x^2 + y^2} = tan^{-1}(y/x)
Rather than trying to solve for y in that equation, I think it would be easier to differentiate implicitly, and then solve for dy/dx in the resulting equation. If you get that far, you want to evaluate the derivative at x = 0, since your polar point (pi, pi/2) has an x-coordinate of 0.