Adding the clinical framing, because it changes how the question reads. Order of operations matters more than any single choice here: establish a baseline, change one thing, wait long enough for it to express itself, then measure again under the same conditions.
Worth separating that from the trial evidence, which this thread keeps folding into the same question. They behave differently and the advice does not transfer.