Background: There is no standardized definition for infant eczema, and various tools have been used across studies, precluding direct comparison.

Objective: The aim of the study was to assess and to compare the accuracy of diagnostic tools for infant eczema using the extensive data collected in Melbourne Infant Study: BCG for Allergy and Infection Reduction (MIS BAIR), an eczema prevention trial.

Methods: Eczema incidence was assessed by 3 questionnaire-based measures: modified UK diagnostic tool, parent-reported medically diagnosed eczema, and parent-reported use of topical corticosteroids. Agreement between the definitions was quantified using κ coefficient. Eczema severity was assessed by 3-monthly Patient-Oriented Eczema Measure (POEM) scores and a SCORing Atopic Dermatitis (SCORAD) clinical assessment at a 12-month visit (ClinicalTrial.gov: NCT01906853).

Results: Among the 538 participants fulfilling at least 1 of the 3 questionnaire-based eczema definitions, only 197 participants (37%) met all 3 definitions. Agreement between the definitions was poor with κ coefficients ranging from -0.11 to 0.62. The most frequently reported symptoms were generally dry skin (483/538, 90%) and pruritus (400/538, 74%). The face (352/538, 65%) and the trunk (306/538, 57%) were more frequently affected than the creases (257/538, 48%). Participants fulfilling all 3 questionnaire-based definitions of eczema were more likely to have higher severity scores and earlier onset of symptoms.

Conclusions: There is poor agreement between currently available tools for assessing infant eczema.