when we determine the last hemoglobin of the month on a group of
patients, then we run a pivot table to determine the distribution of
these hemoglobins. Each person should have one hemoglobin, but
regardless of what identifier we use, MRN, name, etc, invariably there
are a few duplicates in the list. If we count, it doesn't give us the
value of the hemoglobin, there's just a "1", or "2". If we use sum,
the 2s are added and are falsely elevated (8.8+9.0= 17.8), if we
average, the 1s don't show the actual value.

Is there another way to get these values accurately, either by Pivot
Table or some other way?