Validation of five patient‐reported outcomes for atopic dermatitis severity in adults

Abstract

Background Structured patient‐reported outcomes of AD severity are not standardized in clinical practice. We sought to determine the construct validity, internal consistency, cross‐cultural validity, floor or ceiling effects of multiple AD severity assessments. Methods A cross‐sectional, population‐based study of 2893 adults, including 602 adults that met a modified UK Diagnostic Criteria for AD. AD severity was assessed using self‐reported global AD severity, Patient‐Oriented Eczema Measure (POEM), Patient‐Oriented Scoring AD (PO‐SCORAD) and its objective and subjective components, and numerical rating scale (NRS)‐itch. QOL was assessed using short‐form (SF‐)12 mental and physical health scores, SF‐6D health utility scores, Dermatology Life Quality Index (DLQI). Hospital Anxiety and Depression Scale (HADS) was assessed. Results PO‐SCORAD, PO‐SCORAD objective and subjective sub‐scores, NRS‐itch and POEM all had moderate to strong correlations with each other, and DLQI, fair to moderate correlations with HADS‐A and HADS‐D, and inverse correlations with SF‐12 MCS and SF‐6D (Pearson correlations, P<0.0001). All scores showed good criterion validity as judged by analysis of variance and receiver operator characteristics. PO‐SCORAD, PO‐SCORAD objective sub‐scores and POEM had similarly good internal consistency (Cronbach's alpha=0.84, 0.82 and 0.86); PO‐SCORAD subjective sub‐score was less internally consistent (alpha=0.57). All scores showed potentially poor cross‐cultural validity as demonstrated by uniform and non‐uniform differential item functioning by age, sex and/or race/ethnicity for multiple items. There were floor effects for POEM, but not for the other assessments Conclusions PO‐SCORAD, PO‐SCORAD objective and subjective sub‐scores, NRS‐itch and POEM appear to be valid for assessing AD severity in clinical practice.