Unexpectedly, this study did not find evidence for an important role of proprioception in the outcome of exercise therapy. Despite an average improvement in proprioceptive accuracy of 36%, this improvement was not associated with any outcome measures. This suggests that exercise-induced effects are not driven through change in proprioception. Another explanation could be that the measure of knee jointproprioception may not be a valid test. A passive test (i.e. threshold to detection of passive joint motion) in a non- weightbearing position was used in this study.