Reply from the Authors
Peter James Dyck, William J. Litchy, Kenneth A. Lehman, J. L. Hokanson, Phillip A. Low, Peter C. O’Brien · Neurology · 1996
Reply from the Authors: Thank you for agreeing that abnormal limits for neuropathy tests such as attributes of nerve conduction, thresholds of quantitative sensory, or autonomic testing should not be set only by age but by various other physical variables. The approach we recommended for estimating percentiles involved first computing Z scores, measuring how much an observed value differed from the value that would be expected among normal subjects after taking into account these additional variables. As you point out, the percentiles that can be estimated are (appropriately) limited by the sample size used. For example, one cannot estimate a 99.99 centile from a study on 100 subjects. In some applications (research studies, for example), it is desirable to have an ordering beyond the largest available percentiles, and we agree with Redmond and McKenna that the Z …