Type II error in the shoulder and elbow literature.

Abstract

Although the randomized controlled trial (RCT) has been recognized as the gold standard for comparing interventions, trials that fail to show a difference between 2 interventions may suffer from type II error. This study was conducted to determine if the published RCTs in the shoulder and elbow literature have sufficient power to minimize potential type II error and conclude accurately that no difference between interventions exists. We searched Medline for RCTs with negative results in Arthroscopy, the Journal of Bone and Joint Surgery (American), and the Journal of Shoulder and Elbow Surgery (1994 through 2007) with strict inclusion criteria pertaining to care of shoulder and elbow injuries. Eligible studies were analyzed for type II error and the power of their conclusions. The power to detect a difference, if in fact one was present, was only 41% (common standard, 80%). It is a critically important distinction to conclude no difference was observed rather than no true difference between interventions exists.