Background: Reliable assessment of severity in psoriasis is essential to document treatment responses in clinical research. The reliability of current clinical outcome measures is uncertain.
Objective: To quantify the relative variation in commonly used outcome measures (the Psoriasis Area and Severity Index [PASI] and one version of the Psoriasis Global Assessment [PGA]), and a newer measure, the Lattice System Physician's Global Assessment (LS-PGA).
Methods: Physicians who were experienced (53%; 9/17) or inexperienced (47%; 8/17) in using PASI and PGA evaluated 35 patients with psoriasis in random order twice with each rating system. We assessed the variation in scoring psoriasis severity within (intrarater) and among (interrater) physicians.
Results: PASI, PGA, and LS-PGA were highly correlated (r > 0.8 for all comparisons) and had high overall reliability (Cronbach's alpha > 0.9 for each). PGA and LS-PGA had lower intrarater variation than PASI. LS-PGA had a 55% higher concordance coefficient between the two evaluations than did PGA. Interrater variation was lower for PGA and LS-PGA than for PASI both before and after correction for measurement error. Experience was beneficial in reducing variation in PASI scores but was not required with PGA or LS-PGA.
Conclusion: The LS-PGA, which is standardized, does not require experience, and provides discrete word-based scores with intrinsic meaning, is a reliable measure of therapeutic effect in psoriasis, and would allow comparisons across different clinical trials.