Minimal clinically important difference in radiological progression of joint damage. A definition based on patient perspective

Objective. To estimate a threshold for minimal clinically important radiological progression of joint damage using its longitudinal relation with functional disability in patients with rheumatoid arthritis (RA). To validate existing estimates of minimal clinically important difference (MCID) using this relation with functional disability. Methods. We reanalyzed published data of 185 patients with early RA followed for a maximum of 9 years. Longitudinal regression (mixed models) was used, relating radiological damage (modified Sharp score) to functional disability (HAQ-DI), correcting for age (age at diagnosis and increasing disease duration), disease activity (DAS28), and demographic variables. Several shapes of the relation were investigated. Based on the observed relationship between radiological damage, functional disability, and the minimal clinically relevant increase in functional disability found in earlier studies, MCID for progression of joint damage was discussed. Existing estimates of MCID were evaluated for their influence on functional disability over the disease course. Results. A longitudinal relation between the modified Sharp score and the HAQ-DI was found. Significant covariates were age, gender, and disease activity. The model indicated that the relation between the Sharp score and the HAQ-DI was dependent on the amount of damage (a threshold effect) and on patients' age. With lower age, no effect of joint damage on functional disability could be demonstrated and with higher age the effect of joint damage increased. With a typical patient from our cohort (age at diagnosis 55 yrs, some baseline damage, and an expected disease duration of 30 yrs), a (constant) progression of 6 points per year led to an increase of about 0.2 on the HAQ score, solely related to damage, over the disease course. This estimate of MCID was close to estimates based on expert opinion and equal or smaller than most estimates based on the smallest detectable difference. Conclusions. The MCID, defined using longitudinal effect on functional disability, is dependent on age and (progression of) joint damage. However, with a typical patient population this MCID was similar to thresholds based on expert opinion, adding to the validity of these estimates.