Minimal clinically important decline in physical function over one year: EPOSA study

Abstract

Background: The Australian/Canadian hand Osteoarthritis Index (AUSCAN) and the Western Ontario and McMaster Universities knee and hip Osteoarthritis Index (WOMAC) are the most commonly used clinical tools to manage and monitor osteoarthritis (OA). Few studies have as yet reported longitudinal changes in the AUSCAN index regarding the hand. While there are published data regarding WOMAC assessments of the hip and the knee, the two sites have always evaluated separately. The current study therefore sought to determine the minimal clinically important difference (MCID) in decline in the AUSCAN hand and WOMAC hip/knee physical function scores over 1 year using anchor-based and distribution-based methods.

Methods: The study analysed data collected by the European Project on Osteoarthritis, a prospective observational study investigating six adult cohorts with and without OA by evaluating changes in the AUSCAN and WOMAC physical function scores at baseline and 12-18 months later. Pain and stiffness scores, the performance-based grip strength and walking speed and health-related quality of life measures were used as the study's anchors. Receiver operating characteristic curves and distribution-based methods were used to estimate the MCID in the AUSCAN and WOMAC physical function scores; only the data of those participants who possessed paired (baseline and follow up-measures) AUSCAN and WOMAC scores were included in the analysis.

Results: Out of the 1866 participants who were evaluated, 1842 had paired AUSCAN scores and 1845 had paired WOMAC scores. The changes in the AUSCAN physical function score correlated significantly with those in the AUSCAN pain score (r = 0.31). Anchor- and distribution-based approaches converged identifying 4 as the MCID for decline in the AUSCAN hand physical function. Changes in the WOMAC hip/knee physical function score were significantly correlated with changes in both the WOMAC pain score (r = 0.47) and the WOMAC stiffness score (r = 0.35). The different approaches converged identifying two as the MCID for decline in the WOMAC hip/knee physical function.

Conclusions: The most reliable MCID estimates of decline over 1 year in the AUSCAN hand and WOMAC hip/knee physical function scores were 4 and 2 points, respectively.

Keywords: AUSCAN; Functional decline; Hand; Hip/knee; MCID; Osteoarthritis; WOMAC.

Fig. 1

Persons recruited and analysed

Fig. 2

Estimates of MCID percentages for the AUSCAN hand physical function. AUSCAN: Australian/Canadian Hand Osteoarthritis Index; MCID: Minimum Clinically Important Difference; ROC pain: receiver operating characteristic using the AUSCAN pain score as the anchor; J: point that maximizes the Youden index; D: point that minimizes the Euclidean distance; S: point that minimizes the equality sensitivity, specificity. SRM: standardized response mean; SRM2: SRM with Cohen’s threshold 0.20; SRM5: SRM with Cohen’s threshold 0.50; SRM8: SRM with Cohen’s threshold 0.80; CI: Confidence Interval; SEM: Standard Error Measurement; SEM63: SEM with 63% CI; SEM90: SEM with 90% CI; SEM95: SEM with 95% CI; EN: Edwards-Nunnally index; EN90: EN with 90% CI; EN95: EN with 95% CI

Fig. 3

Estimates of MCID percentages for the WOMAC hip/knee physical function. WOMAC, Western Ontario and McMaster Universities; MCID: Minimum Clinically Important Difference; ROC pain: receiver operating characteristic using the WOMAC pain score as the anchor; ROC stiffness: receiver operating characteristic using the WOMAC stiffness score as the anchor; J: point that maximizes the Youden index; D: point that minimizes the Euclidean distance; S: point that minimizes the equality sensitivity, specificity. SRM: standardized response mean; SRM2: SRM with Cohen’s threshold 0.20; SRM5: SRM with Cohen’s threshold 0.50; SRM8: SRM with Cohen’s threshold 0.80; CI: Confidence Interval; SEM: Standard Error Measurement; SEM63: SEM with 63% CI; SEM90: SEM with 90% CI; SEM95: SEM with 95% CI; EN: Edwards-Nunnally index; EN90: EN with 90% CI; EN95: EN with 95% CI