As such,the SD just tells us the distribution of the recent group to which you are being compared. It is (largely) irrelevant when comparing two candidates, unless you want to get a rough idea of the percentile their score represents.
As aPD mentioned, the standard arror is important, but this is a bit different from the standard error of a sample of measurements etc etc. As aPD said, this represents the interval around your score for which the USMLE-peeps are 95% confident you belong.
If the two intervals of candidates overlap, THEN there is no real difference between their results. If they fall within the same standard deviation from the mean, it does NOT mean that they are not significantly different in their ability to perform on the USMLE!
This old post is helpful:
http://mydominanthemisphere.wordpress.com/2008/05/06/usmle-scores-debunking-common-myths/