A model of impairment and functional limitation in rheumatoid arthritis

Abstract

Background: We have previously proposed a theoretical model for studying physical disability and other outcomes in rheumatoid arthritis (RA). The purpose of this paper is to test a model of impairment and functional limitation in (RA), using empirical data from a sample of RA patients. We based the model on the disablement process framework.

Methods: We posited two distinct types of impairment in RA: 1) Joint inflammation, measured by the tender, painful and swollen joint counts; and 2) Joint deformity, measured by the deformed joint count. We hypothesized direct paths from the two impairments to functional limitation, measured by the shirt-button speed, grip strength and walking velocity. We used structural equation modeling to test the hypothetical relationships, using empirical data from a sample of RA patients recruited from six rheumatology clinics.

Results: The RA sample was comprised of 779 RA patients. In the structural equation model, the joint inflammation impairment displayed a strong significant path toward the measured variables of joint pain, tenderness and swelling (standardized regression coefficients 0.758, 0.872 and 0.512, P <or= 0.001 for each). The joint deformity impairment likewise displayed significant paths toward the measured upper limb, lower limb, and other deformed joint counts (standardized regression coefficients 0.849, 0.785, 0.308, P <or= 0.001 for each). Both the joint inflammation and joint deformity impairments displayed strong direct paths toward functional limitation (standardized regression coefficients of -0.576 and -0.564, respectively, P <or= 0.001 for each), and explained 65% of its variance. Model fit to data was fair to good, as evidenced by a comparative fit index of 0.975, and the root mean square error of approximation = 0.058.

Conclusion: This evidence supports the occurrence of two distinct impairments in RA, joint inflammation and joint deformity, that together, contribute strongly to functional limitations in this disease. These findings may have implications for investigators aiming to measure outcome in RA.

Figure 1

Identification diagram of a structural equation model of the relationship between the stages of impairment and functional limitation in rheumatoid arthritis. Two types of impairment, joint inflammation and joint deformity are shown as ovals on the left. Measurements for these latent variables include joint tenderness (JT), joint swelling (JS) and joint pain (JP), for joint inflammation; and joint deformities. We disaggregated joint deformities into upper limb (DUL), lower limb (DLL) joints, and other joints (DOJ). Several of the parameters were constrained to enable estimation. Circles represent residuals or disturbance terms, for each variable. See Table 1 for parameter estimates.

Figure 2

Frequency distributions of the joint inflammation (JI) impairment latent variable. This was computed from JI = JT^1/2 + JS^1/2 × 0.445 + JP^1/2 × 0.837, where JT = joint tenderness, JS = joint swelling, and JP = joint pain. Weights were estimated using maximum likelihood with Amos, after constraining the coefficients for JT and DUL to 1. The latent variable was then rescaled to vary from 0 to 100.

Figure 3

Frequency distributions of the joint deformity (JD) impairment latent variable. This was computed from JD = DUL + DLL × 0.478 + DOJ × 0.011, where DUL = deformity upper limb, DLL = deformity lower limb, and DOJ = deformity other joints. The weights for the equation were estimated using maximum likelihood with Amos, after constraining the coefficients for JT and DUL to 1. The latent variable was then rescaled to vary from 0 to 100.