Background/Purpose: Use of PROs to assess health-related quality of life in clinical practice, research studies, and clinical trials in Rheumatoid Arthritis (RA) remains an ongoing area of research. SF36 is commonly used in RA trials but is not feasible for routine use in clinical practice settings. The PROMIS (Patient Reported Outcomes Measurement Information System) may address this gap but has not been widely assessed in RA patients starting therapy in a real-world comparative effectiveness study, nor examined in that setting in relation to the SF36 and Clinical Disease Activity Index (CDAI). These were evaluated in AWARE (Comparative and Pragmatic Study of Golimumab IV Versus Infliximab in Rheumatoid Arthritis), an ongoing Phase 4 study providing real-world assessment of IV Tumor Necrosis Factor inhibitor (TNFi) medications in RA pts.

Methods: AWARE is a prospective, noninterventional, 3-year study at 88 US sites. RA pts were enrolled when initiating TNFi treatment. Treatment decisions were made by treating rheumatologists. We report baseline PROMIS-29 (7 domains and pain intensity), PROMIS Pain Interference (PI) Short Form (SF) 6b (PI6b) and PROMIS Fatigue (F) Short Form 7a (F7a), domain T-Scores, and SF36 subdomain and Component Scores (CS) in AWARE patients. Here we report baseline data obtained from the final 1-year AWARE dataset. Correlations between PROMIS measures and comparable SF36 component scores were calculated using Pearson Correlations. Data is shown as mean ± standard deviation (SD).

Results: At baseline, mean CDAI of all pts (n=1262) was 32.3±15.6, with 70.4% in High Disease Activity (HDA, CDAI >22), 22.8% in Moderate DA ( >10 and ≤22), 6.1% in Low DA ( >2.8 and ≤10) and 0.7% in Remission (≤2.8). Mean PROMIS scores were >0.5 SD worse than population means for Physical Function (PF, 38.1±6.84), PI (63.4±7.68), F (58.8±9.95), Sleep Disturbance (55.1±8.68); and Ability to Participate in Social Roles/Activities (PSRA, 43.4±8.58). Baseline Depression and Anxiety were within 0.5 SDs of population T-scores.  PI6b, F7a, and P29 domain T-scores correlated with the comparable SF36 subdomain and component scores (r’s >0.58), excepting sleep for which no comparable SF36 element was applicable. Examples include: P6b (r=-0.80) and P29-PI (0.81) with SF-36 Bodily Pain; F7a (-0.77) and P29-F (-0.77) with SF36 Vitality; P29-PF with SF36 PF (0.77), Role-Physical (0.69), and Physical CS (0.73); P29 Anxiety with SF36 Mental Health (-0.72), Role- Emotional (-0.56), Mental CS (-0.70); and P29-PRSA with SF36-Social Functioning (0.71). Mean PROMIS-29 T-scores (except Anxiety and Sleep Disturbance) among patients with HDA were significantly different from patients with MDA, LDA or remission (p < 0.001 for all).  Further, mean PROMIS T-scores of PF, F, PSRA, PI, Pain Intensity, PI6b and P7a among patients with MDA were significantly different from patients with more or less active RA (by CDAI category).

Conclusion: Analysis of baseline results from a large cohort of RA patients indicates high correlations between individual P29 domain T-scores and SF36 component scores, as well as categorical CDAI, providing strong evidence of PROMIS construct validity in a real-world population of RA patients.

« Back to 2019 ACR/ARP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/construct-validation-of-promis-short-form-and-profile-29-t-scores-with-sf-36-in-rheumatoid-arthritis-patients-treated-for-1-year-results-from-a-real-world-evidence-based-study-in-the-united-states/