[Skip to Content]
[Skip to Content Landing]
Citations 0
July 14, 1999

Gabapentin for Painful Diabetic Neuropathy—Reply

JAMA. 1999;282(2):133-134. doi:10-1001/pubs.JAMA-ISSN-0098-7484-282-2-jbk0714

In Reply: The scale used in our study, to which Drs Schiebel and Ebbert refer, was not a true Likert scale as it did not have a pain description associated with each number. Only the extreme ends of the scale had word descriptions (0=no pain, 10=worst possible pain) so our scale could be more accurately characterized as a numerical rating scale. Numerical rating scale scores can be analyzed using parametric methods.1 Furthermore, we did not analyze single numerical rating scale scores, but rather the mean of daily scores recorded over a week for each patient. The treatment group mean was an average of the means per patient. Since the central limit theorem2 holds that means tend to be normally distributed, it was appropriate to use parametric analyses.