Challenges in dealing with walking speed in knee osteoarthritis gait analyses

Background: The knee osteoarthritis biomechanics literature is varied in how researchers deal with the potentially confounding effects of differences in walking speed between individuals with osteoarthritis and healthy controls, and between increasing levels of knee osteoarthritis severity. While many studies choose to have participants walk at a self-selected pace in an attempt to observe and capture natural joint mechanics, some have chosen to have individuals walk at various speeds and analyze trials at pre-determined speeds. Others have chosen to statistically control for the confounding effects of speed by using an analysis of covariance model. This has made it difficult to assimilate findings of various studies into a consensus of biomechanical changes potentially associated with disease initiation and progression. And it begs the question of which method(s) are correct. Methods: A discussion around this point is overdue and this report is an attempt to begin this discussion using some examples from the knee osteoarthritis literature. Findings and Interpretation: Key conclusions of this report include the inappropriateness of using an analysis of covariance to statistically control for speed in studies of knee osteoarthritis, and the need to use caution when interpreting the results of studies that control for speed in different ways.