Background/Purpose:  The Psoriasis Symptom Inventory (PSI) is an 8-item patient-reported outcome measure of psoriasis symptom severity. Data from a Phase 2 study of Brodalumab in subjects with plaque psoriasis demonstrated that the PSI has good reliability, validity, and ability to detect change. Psoriasis related symptoms are an important component of psoriatic arthritis (PsA). However, the measurement properties of the PSI have not been evaluated in the PsA population. This analysis sought to evaluate the reliability and construct validity of the PSI in subjects with PsA.

Methods: This was a secondary analysis of pooled data from treatment arms of a Phase 2 clinical trial (NCT01516957) evaluating the efficacy of Brodalumab, an anti-IL-17 R monoclonal antibody in subjects with PsA. Confirmatory factor analysis (CFA) and Rasch analysis were used to assess the dimensionality of the PSI. Item evaluation and internal consistency (Cronbach’s α) were conducted on baseline PSI data. Test-retest reliability was assessed using intraclass correlation coefficients (ICC) between PSI scores at week 2 and week 4 in stable subjects (i.e., -1≤ change ≤1 on the subject global assessment of disease [SGA]). Construct validity was evaluated based on correlations between PSI scores and body surface area (BSA) affected by psoriasis, and selected domains of the SF-36. Known groups validity was explored based on BSA severity categories (<5%, 5-10%, >10%) using analysis of variance. Ability to detect change was explored using t-tests comparing mean PSI scores in subjects reporting ≥30% versus <30% improvement from baseline to week 12 on the SGA.  

Results: The analysis sample included 154 subjects; 93.5% White, 63.0% females, mean (SD) age was 52.2 (11.47) years. Mean (SD) duration of PsA and BSA at baseline was 8.8 (7.84) years and 10.4% (15.61%) respectively. At baseline, 12% of subjects had no skin involvement and 63% had ≤5% skin involvement. Mean (SD) PSI total score at baseline was 12.2 (7.89). CFA and Rasch analysis supported unidimensionality. Rasch analysis also indicated good item fit and correctly ordered categories. The PSI had excellent internal consistency (α=0.95) and good test-retest reliability (ICC=0.70 for total scores and ranging from 0.67 to 0.81 for items). Convergent validity was supported by moderate correlations with BSA (r=0.50). Discriminant validity was supported by small correlations (r<-0.3) for SF-36 domains of mental health and role emotions. Known groups validity was supported by significantly lower mean PSI scores (p<0.001) between subjects with BSA<5% compared to those with BSA>10%). Mean change in PSI score was significantly greater (p<0.001) in subjects with ≥30% SGA improvement than subjects with <30% SGA improvement.  

Conclusion: This study provides evidence that the PSI is unidimensional, with excellent internal consistency, good test-retest reliability, construct validity, and ability to detect change in subjects with PsA. Based on the findings, the PSI is a robust yet simple and practical measure of psoriasis-related symptoms for use in PsA clinical trials.

---

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/reliability-and-construct-validity-of-the-psoriasis-symptom-inventory-in-subjects-with-psoriatic-arthritis/