The objective is to evaluate and diagnose, in a controlled setting, suspected food allergy causation in patients hospitalized for management of severe, unremitting atopic dermatitis (AD). Nineteen children were hospitalized at Oregon Health and Science University with atopic dermatitis from 1986 to 2003 for food restriction, then challenge, following standard recommendations. Challenges were prioritized by categories of (a) critical foods (e.g., milk, wheat, egg, soy); (b) important foods; and (c) other suspected foods. Patients were closely observed for evidence of pruritus, eczematous responses, or IgE-mediated reactions. If results were inconsistent, double-blind, placebo-controlled food challenge was performed. A total of 17 children with atopic dermatitis were assessed. Two could not be fully evaluated, thus were excluded from data tabulations. Only one positive eczematous food response was observed of 58 challenges. Three children had well-documented histories of food-induced IgE-mediated anaphylactoid or urticaria reactions to seafood and/or nuts and were not challenged with those foods. Atopic dermatitis, even in the highest-risk patients, is rarely induced by foods. Undocumented assumptions of food causation detract from proper anti-inflammatory management and should be discouraged. Immediate IgE-mediated food reactions are common in atopic dermatitis patients; such reactions are rapid onset, typically detected outside the clinic, and must be distinguished from eczematous reactions. Diagnosis of food-induced eczema cannot be made without food challenge testing. Such tests can be practical and useful for dispelling unrealistic assumptions about food allergy causation of atopic dermatitis.