Predictors of Response to Physical Therapy Intervention in Patients With Primary Hip Osteoarthritis

Abstract

Few studies have investigated or identified common clinical tests and measures as being associated with progression of hip osteoarthritis (OA); fewer still are longitudinal studies exploring prognostic variables associated with long-term outcome following physical therapy treatment. The purpose of this study was to determine a set of prognostic factors that maximize the accuracy of identifying patients with hip osteoarthritis (OA) likely to demonstrate a favorable response to physical therapy intervention. This was a prognostic study. Ninety-one patients with a clinical diagnosis of hip OA were analyzed to determine which clinical measures, when clustered together, were most predictive of a favorable response to physical therapy intervention. Responders were determined based on OMERACT-OARSI response criteria, which included percent and absolute changes in pain, function, and global rating of change over 1 year. These data served as the reference standard for determining the predictive validity of baseline clinical examination variables. Using multivariate regression analyses and calculations for sensitivity, specificity, and positive and negative likelihood ratios, a cluster was identified. Five baseline variables (unilateral hip pain, age of ≤58 years, pain of ≥6/10 on a numeric pain rating scale, 40-m self-paced walk test time of ≤25.9 seconds, and duration of symptoms of ≤1 year) were retained in the final model. Failure to exhibit a condition of 1 of the 5 predictor variables decreased the posttest probability of responding favorably to physical therapy intervention from 32% to <1% (negative likelihood ratio=0.00, 95% confidence interval=0.00-0.70). Having at least 2 out of 5 predictor variables at baseline increased the posttest probability of success with physical therapy intervention from 32% to 65% (positive likelihood ratio=3.99, 95% confidence interval=2.66-4.48), and having 3 or more of 5 predictor variables increased the posttest probability of success to 99% or higher. A comparison with a control group that did not receive physical therapy further substantiated the cluster. The small sample size and the number of variables entered into the logistic regression model may have resulted in spurious findings. This study must be validated in replication studies before it can be considered for use in clinical practice. This study completed the first step in the development of a preliminary cluster of baseline variables that identify patients with hip OA as positive responders to physical therapy intervention.