Re-evaluating "Success" as It Pertains to Surgical Trials.

The objective of this article is to review the difficulties with the design and interpretation of surgical clinical trials. Few surgical procedures are evaluated in a randomized fashion. There are a number of factors that make the design of surgical trials diffiuclt, and many surgical questions cannot be answered with a clinical trial. Issues with standardization of the surgical procedure, variability of surgical skills, and changes in surgical expertise over time further complicate the design and implementation of surgical trials. Statistical methods for surgical trials often require a noninferiorty design and are more complicated to interpret than the more common superiority trial. Even when properly conducted, both superiority and noninferiority trials are often misinterpreted. Because of the relatively high success rate in surgery, trials require large numbers of patients and noninferiority trials are often inconclusive with respect to the primary outcome. Surgical trials are often misinterpreted or over interpreted, and there can be confusion in how the findings of these trials should be incorportated into clinical practice. The interpretation of the results of a surgical trial often differ significantly from the primary and secondary outcomes that were specified in the trial design.