We determined sample sizes for chi-square testing based on a corticosteroid injection failure rate of 30% and a 20% difference in grip strength (flexion vs extension) in symptomatic lateral epicondylitis.2,14 Power calculations were done for the recurrence of lateral epicondylitis. With this estimate and 25 subjects in each of the 3 groups, a recurrence rate of 30% or lower could be detected for the iontophoresis group, assuming an of 0.05, power of 80%, and 2-sided testing. For the continuous measurement with 25 subjects per group, a mean difference between the 3 treatment groups of 0.63 or larger standard deviation units are detectable.