Short answer first, then the reasoning. The single most useful discipline is bringing the whole panel to a clinician rather than one flagged value. One out-of-range result in isolation generates anxiety and unnecessary tests; the same result next to the trend and the rest of the panel usually generates a shrug.
That is the short version; the long version is somebody else's post.