Averaging the results? They probably discard outliers too. Unusually high output is something you'd want to be aware of rather than hiding it in an averaged result.
Is it OK to give extremely high doses to, say, 1 in 100 people, if the other 99 are normal?