Self-reported physical functioning was more influenced by pain than performance-based physical functioning in knee-osteoarthritis patients

Abstract

Background and Objectives

To test the hypothesis that self-reported physical functioning is more influenced by pain than performance-based physical functioning.

Methods

163 knee-osteoarthritis patients completed the performance-based DynaPort^® KneeTest (DPKT), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), and SF-36 (self-report measures of pain and physical functioning) before, 3, 6, and 12 months after knee replacement.

Results

Correlations between (two) self-reported measures of functioning and (two) pain measures were higher (0.57–0.74) than correlations between the performance-based measure of functioning and the two pain measures (0.20 and 0.26). In factor analysis, WOMAC and SF-36 pain and physical functioning subscores loaded on the first factor (eigenvalue 3.2), while DPKT KneeScore2 loaded on the second factor (eigenvalue 0.92). Before surgery, correlations between performance-based and self-reported physical functioning were higher in patients with less pain (0.43) compared to patients with more pain (0.17), for the WOMAC (as expected), but not for the SF-36. After surgery, when the pain had diminished, the correlations between performance-based and self-reported physical functioning were higher, especially for the WOMAC.

Conclusions

Our hypothesis was convincingly supported by the results of the WOMAC, and somewhat less by the results of the SF-36. We consider this as evidence for a lack of content validity of the WOMAC.

Introduction

There is an ongoing debate about the validity of performance-based measures vs. self-report questionnaires to assess physical functioning in patients with hip and knee osteoarthritis (OA) [1], [2], [3], [4], [5], [6], [7], [8], [9]. A performance-based measure is one in which an individual is asked to perform one or more specific tasks that are evaluated in a standardized manner using predefined criteria, such as counting repetitions or timing of the activities [10]. A self-report measure is one in which an individual is asked to indicate his/her perceived level of functioning during daily activities, described in standardized questions.

A number of arguments have been proposed in favor of one or the other method. Some authors prefer self-report questionnaires because these are claimed to be easier to use, less time-consuming, and less of a burden to patients [8]. Others argue that self-report questionnaires are preferred because they cannot be influenced by observer bias [5]. Performance-based methods have been considered less valid because they measure physical functioning in an artificial situation, are influenced by the subject's motivation to participate, and may provide little information about how a person copes in his/her own environment [6], [7]. On the other hand, performance-based methods are claimed to be less influenced by psychologic factors such as expectations and beliefs [9], cognitive impairments [4], culture, language, and education level [1], [7]. Furthermore, it has been suggested that performance-based measures may identify early deficits in physical functioning (“preclinical disability”) before they are identified by self-reports [2], [3].

There is a paucity of studies that empirically investigate these claims. Numerous studies report moderate correlations between performance-based measures and self-report measures [11], [12], [13], [14], [15], but most hypotheses about possible explanations have not been empirically tested.

Recently Stratford et al. [13] suggested two possible explanations for the modest correlation between performance-based measures and self-report measures of physical functioning in 93 patients awaiting hip or knee arthroplasty. These explanations were: (1) lack of reliability of the performance-based measures; and (2) lack of content validity of the performance-based measures. They argued that performance-based measures based on time alone inadequately represent the breadth of health concepts associated with functional status. In a study in 93 patients awaiting total hip or knee replacement they found a moderate correlation of a self-report measure (the Lower Extremity Functional Scale—LEFS) with a performance-based measure of physical functioning (40-meter fast self-paced walk). To increase the reliability of the performance-based measure, three performance-based functioning scores (self-paced walk, timed up-and-go, and stair test) were summed. In contrast with their hypothesis, the correlation of the LEFS with the sum of the three timed scores was not higher than the correlations of the LEFS with the individual timed scores. They concluded that measurement error could not explain the modest correlation between self-report measures of physical functioning and performance-based measures of physical functioning. In contrast, by adding scores for pain and exertion to the performance-based time score, the correlation with self-reported physical functioning increased. They considered this as evidence for a lack of content validity of the performance-based test [13].

However, based on their results, we would draw the opposite conclusion, and consider their results as evidence for a lack of content validity of the self-report measures of physical functioning. Stratford et al.'s results show that self-reported levels of functioning refer not only to the time to complete the task, but also to pain and exertion, while these measures claim to measure only functioning, not a combination of pain and functioning. Many self-report questionnaires, like the Western Ontario and McMaster University Osteoarthritis Index (WOMAC) [16] and the MOS 36-item Short Form Health Survey (SF-36) [17], claim to measure pain and functioning as two different constructs, with different subscales.

Following this line of reasoning, our hypothesis is that self-report measures of physical functioning will be more influenced by the amount of pain experienced than performance-based measures of physical functioning. As a consequence, we expect that the correlation between self-report measures of physical functioning and performance-based measures of physical functioning is higher in patients with less pain. The aim of this study was to test this hypothesis in a population of patients undergoing knee replacement surgery.