Responsiveness and Minimally Clinically Important Differences of PROMIS Measures in Rheumatoid Arthritis

Background/Purpose: The ability to detect meaningful change in clinical status (responsiveness) is an important aspect of validity. Minimally clinically important difference (MCID) is a patient-centered construct reflecting the smallest difference of value to patients. PROMIS is a generic family of measures developed for use across chronic diseases. We examined the responsiveness and estimated MCIDs of selected PROMIS measures reflecting domains that people with RA have identified. 

Methods: Data are from the first two visits of an observational trial in Baltimore, MD. Patients completed PROMIS Physical Function (PF), Pain Interference (PI), Participation Ability (PA), and Fatigue computer adapted tests and other patient-reported outcomes (PROs) using a tablet. At the second visit, patients also completed a 5-point item assessing change in RA status from the previous visit (much worse to much better) and the self-reported health rating. Descriptive statistics were calculated, and ANOVA was used to identify significant differences in scores on 9 PROMIS measures.

Results:  The 196 RA patients who completed outcomes at clinical visits approximately 4 months apart were mostly female (81%), white (82%) with a mean age of 55 (13). All met ACR 1987 or 2010 criteria for RA.  One-third reported the same health status at both visits (68 [35%]); 13% were a little better; a similar number were much better; 27% were a little worse, and 7% were much worse. Correlations between PROMIS and other measures assessing similar domains ranged from 0.32 to 0.83. Among patients reporting being “much worse”, PROMIS scores worsened from 1.0 to 8.1 points (mean CDAI change 9.2)(Table). Among patients feeling much better, PROMIS scores improved from 1.6 to 6.9 points (mean CDAI change -6.6). Change in PROMIS scores were largest in relation to changing disease activity for symptoms highly relevant to RA (e.g., PI, fatigue, PF); notably patients with worsening in RA did not report higher anxiety or depression scores. 

Conclusion: These initial data suggest PROMIS measures are responsive to clinical changes in RA status and contribute to the growing literature supporting the use of PROMIS measures to assess physical, social, emotional and health in people with RA.