—We disagree with Dr Gordon that the statistical question asked in clinical testing of a less costly treatment is different from that posed in any other randomized trial. The major hypothesis to be tested in this or any randomized trial is whether the two treatments are different. The interpretation of the results should not be affected by availability or cost of treatment. If a significant difference (P≤.05) is found between two therapies, the conclusion of the study is usually accepted.When, as in our study, treatment groups are not statistically different with regard to a chosen end point (death), some investigators do try to indicate whether the chosen sample size was adequate. However, we do consider a retrospective power analysis to estimate the probability of missing a difference that truly exists to be particularly useful. The power curve is indeed useful and appropriate as Freimen et al
Baughman KL, Kallman CH. Minimizing Medical Costs-Reply. Arch Intern Med. 1983;143(4):845. doi:10.1001/archinte.1983.00350040235050
Artificial Intelligence Resource Center
Customize your JAMA Network experience by selecting one or more topics from the list below.