Content and construct validity, predictors, and distribution of self-reported atopic dermatitis severity in US adults

Atopic dermatitis (AD) is associated with skin lesions, multiple symptoms, and effect of quality of life, all of which factor into disease severity. Self-reported global AD severity may be a valid severity assessment for epidemiologic research.

To validate self-reported global AD severity in a representative cohort of adults with AD.

Preliminary probing-cognitive interviews were performed (n = 8). Next, a cross-sectional US population-based survey study of adults with AD was performed. AD was diagnosed using an adap/tation of the UK Working Party criteria (n = 602). AD severity was assessed using self-reported global AD severity (mild, moderate, severe), Patient-Oriented Scoring AD (PO-SCORAD), Patient-Oriented Eczema Measure (POEM), Numeric Rating Scale (NRS)-itch, NRS-sleep, NRS-pain, and Hospital Anxiety and Depression Scale (HADS).

Self-reported global AD severity had good content validity. Self-reported global AD severity had strong correlations with PO-SCORAD (Spearman correlation ρ = 0.61) and objective PO-SCORAD (ρ = 0.61); moderate correlations with POEM (ρ = 0.54), NRS-itch (ρ = 0.44), NRS-pain (ρ = 0.46), and HADS (ρ = 0.41); and weak correlation with NRS-sleep (ρ = .32) (P < .001 for all). Consistent and significant correlations were observed in stratified analyses by age, sex, race/ethnicity, and level of education. There were stepwise increases of PO-SCORAD, NRS-itch, NRS-sleep, NRS-pain, POEM, and HADS with increasing self-reported global AD severity (Kruskal-Wallis test, P < .01). There was weak-moderate concordance between self-reported AD severity and established severity strata for PO-SCORAD (ρ = 0.44), NRS-itch (ρ = 0.30), and POEM (ρ = 0.43). Rather, self-reported global AD severity was best predicted by a combination of PO-SCORAD, POEM, NRS-itch, NRS-pain, and HADS. No differential item reporting was found by age, sex, or race/ethnicity.

Self-reported AD severity simultaneously assesses multiple AD constructs and appears to be sufficiently valid for assessing AD severity in clinical and epidemiologic studies.