Background While the PROMIS (Patient-Reported Outcomes Measurement Information System) physical function short form 10a (PF10a) is both practical and acceptable for implementation in routine clinical practice, its psychometric properties have not been evaluated in Systemic Lupus Erythematosus (SLE). We examined the validity and responsiveness of PF10a in SLE among a racially/ethnically diverse clinic population and developed estimates of the minimally important difference (MID).

Methods Data were derived from electronic health records for all SLE patients seen in a university-based rheumatology clinic between 2013 and 2018. We evaluated the PF10as floor and ceiling effects among different racial/ethnic groups. Construct validity was assessed by examining Spearmans correlation coefficients between the PF10a and other patient-reported (pain (scale 0–10) and pain visual analogue scale (VAS) (scale 0–100)), physician-reported (SLE disease activity index (SLEDAI)) and laboratory (erythrocyte sedimentation rate (ESR)) measures. Known-group validity was assessed by evaluating effect size (Cohens d) between categories of pain (no pain vs. moderate-severe pain). We used standardized response means to examine the responsiveness of the PF10a to longitudinal changes in pain and SLEDAI. MID was estimated using distribution based and anchor-based methods.

Results We included 612 patients in cross-sectional analyses of validity and 462 patients in longitudinal analyses of responsiveness. Mean age was 40.5±14.6, 87% were female and 32% Caucasian. The PF10a had ceiling effects above the commonly accepted criteria of 15% among Caucasian (23%), Asian (23%) and Other (17%) race/ethnicities, and no floor effects. Construct validity analyses showed strong correlations (=0.66, p<0.05) with pain VAS, moderate correlations (=0.58, p<0.05) with pain, and weak correlations with ESR (=0.25, p<0.05) and SLEDAI (=0.16, NS). Known-group validity analyses showed large differences among pain groups (Cohens d=1.49, p<0.05). The PF10a was responsive to improvements in pain (SRM=0.5) and SLEDAI (0.49), but less so to deteriorations in pain (SRM=−0.42) or SLEDAI (SRM=−0.24). Distribution-based MIDs were +8 for improvement and −7 for deterioration. Anchor-based MIDs were +2 for improvement, −3 for deterioration with pain as anchor and +5 for improvement, −5 for deterioration with SLEDAI as anchor.

Conclusions Although the PF10a showed some ceiling effects, it had good validity in this young racially/ethnically diverse sample with SLE. The PF10a was responsive to improvements in pain and disease activity. The anchor-based MIDs appear to be similar to those reported for PF10a in rheumatoid arthritis. This information supports the use of the PF10a in SLE and provides important information to facilitate interpretation of scores.

Funding Source(s): UCSF PREMIER