Do you routinely check an Allen's test before placing a radial arterial line?

If abnormal, does that data motivate you to cannulate an otherwise less ideal site like the brachial, axillary or femoral? I have encountered varied, strong opinions on this question in different practice settings and from different generations of providers. My understanding is that the Allen's test has not been shown to be a reliable predictor for complications in the radial access cardiac catheterization setting. I have not been able to find good data assessing its utility in critical care.