Invited Commentary

performance-based tests. Many clinicians and investigators would say that total hip and knee arthroplasty are 2 of the most successful surgeries in the field of orthopedics. Much of the justification for such labeling stems from self-reported instruments that can be heavily swayed by the powerful capacity for arthroplasty to reduce pain. These 2 surgeries may represent some of the most extreme and consistent reductions in pain in today’s health care environment. Physical therapy is well positioned as a profession to address the common and often lingering impairments in muscular performance that have consistently been related to residual locomotor deficits. Clinicians may miss opportunities to maximize their patients’ postoperative gains if they fail to take a more comprehensive view of physical function by adding performance-based measures to their outcomes array. I look forward to reading future work on this topic from Stratford and colleagues as they continue to pursue this line of research.

can overestimate the scores on performance-based tests. Many clinicians and investigators would say that total hip and knee arthroplasty are 2 of the most successful surgeries in the field of orthopedics. Much of the justification for such labeling stems from self-reported instruments that can be heavily swayed by the powerful capacity for arthroplasty to reduce pain. These 2 surgeries may represent some of the most extreme and consistent reductions in pain in today's health care environment. Physical therapy is well positioned as a profession to address the common and often lingering impairments in muscular performance that have consistently been related to residual locomotor deficits. Clinicians may miss opportunities to maximize their patients' postoperative gains if they fail to take a more comprehensive view of physical function by adding performance-based measures to their outcomes array. I look forward to reading future work on this topic from Stratford and colleagues as they continue to pursue this line of research.

Invited Commentary
Yong Hao Pua, Kim L. Bennell Much has been written in the literature about the close association between self-report measures of physical function and pain in patients with upper-extremity 1,2 and lowerextremity 3-6 musculoskeletal conditions. That pain wields a strong influence on self-report measures of physical function is neither a trivial nor a purely academic finding: there are instances where pain and physical function can uncouple, which, in turn, limits the ability of self-report measures to assess functional ability. Specifically, in the context of patients who have recently undergone a knee or hip arthroplasty, at least 2 studies 5,6 have shown that these patients "felt better but did worse": that is, self-report measures of pain and physical function improved significantly, but walking and stairclimbing performance was no better-or even worse-than the prearthroplasty performance.
Against this background, the article by Stratford and colleagues 7 provides thought-provoking data on the extent to which self-report measures of function-namely,

Self-Report Measures of Mobility Postarthroplasty
and that of Stratford et al 8 suggest that the various performance measures were differentially correlated with self-report measures of pain and physical function in patients with hip or knee osteoarthritis. Accordingly, to the extent that the fixed biases reported by Stratford et al 7 remained stable across the different performance measures, greater confidence may be placed on the robustness and generalizability of the data.
A second issue relates to the conservativeness of the criteria used by Stratford et al 7 to ascertain whether the observed fixed biases were clinically important. In the first place, it has been argued that that the minimal important difference (also known as the minimal clinically important difference) for a group of patients is lower than that for an individual patient, 9 whereas the minimal important difference (based on a small change) for an individual patient is reportedly lower than the minimal detectable change. 10 -13 Stratford and colleagues' criteria were 9 points for both the WOMAC-PF and the LEFS, which represented the minimal detectable change for the individual patient. 4 Accordingly, it is likely that Stratford and colleagues' yardstick was too stringent when appraising withingroup differences, and, therefore, the authors may have understated the significance of their findings. Finally, so that readers can better appreciate the clinical significance of Stratford and colleagues' findings, it may be useful for the authors to report the proportion of patients whose systematic biases exceeded the minimal important difference 14,15 To sum up, the article by Stratford et al 7 reminds us about the substantial mismatch between self-report and performance measures of physical function in patients who recently underwent a knee or hip arthroplasty. Although Stratford et al aim to provide readers with useful data to appraise past research work, they also should be thanked for providing a compelling argument for the inclusion of performance measures of function in present and future research efforts so that we can draw a more comprehensive picture of the patient's lower-extremity functional status.