That is good advice. However, if the problem is preanalytical, it will likely be positive if retested.

I was not given a chance to retest the same sample, so that datapoint is lost.

So if a retest is positive, it still does not confirm drinking. It seems another type of test altogether, CDT maybe, starts to paint a better picture. Probably also good is an assessment by a shrink.