Background: The purpose of this investigation was to undertake a hypothesis generating study to identify candidate variables that characterize people with knee osteoarthritis who are most likely to experience a positive response to exercise.

Methods: One hundred and fifty participants with knee osteoarthritis participated in this observational, longitudinal study. All participants received a standard exercise intervention that consisted of 20-min sessions two to three times a week for three months. The classification and regression tree methodology (CART) was used to develop prediction of positive clinical outcome. Positive pain and disability outcomes (dependent variables) were defined as an improvement in pain intensity by >50% or an improvement of five or more on the Oxford knee score, respectively. The predictor variables considered included age, sex, body mass index, knee osteoarthritis severity (Kellgren/Lawrence grade), pain duration, use of medication, range of knee motion, pain catastrophizing, self-efficacy and knee self-perception.

Results: Fifty-five participants (36.6%) were classified as responders for pain intensity and 36.6% were classified as responders for disability. The CART model identified impairments in knee self-perception and knee osteoarthritis severity as the discriminators for pain intensity reduction following exercise. No variables predicted reduction of disability level following exercise.

Conclusions: Such findings suggest that both body perception and osteoarthritis severity may play a role in treatment outcome with exercise. It also raises the possibility that those with higher levels of disrupted body perception may need additional treatment targeted at restoring body perception prior to undertaking exercise. Significance: Regardless age, sex, body mass index, pain duration, use of medication, knee range of motion, pain catastrophizing and self-efficacy, participants with knee osteoarthritis who report low levels of body perception disruption (a FreKAQ score ≦ 17) and minimal structural changes (KL grade I) demonstrate significantly better outcomes from exercise therapy than other participants.


hypothesis-generating study, knee osteoarthritis, physical therapy, exercise, treatment outcomes

Link to Publisher Version (URL)


Find in your library