The Development And Validation Of Food Allergy Data Dictionary

Lack of a comprehensive and widely adopted terminology for food allergy (FA) presents a challenge for researchers and clinicians to improve care and outcomes. There is a discernible need to develop a core lexicon and ontology for FA concepts which would enable data sharing across multiple institutions to advance FA research and treatment. During the first phase, clinical artifacts, including, patient intake forms, templates, clinical notes, and pre-compiled EPIC phrases used across four Food Allergy Research and Education (FARE) Clinical Network (FCN) sites were analyzed and interviews with domain experts were conducted. Next, we defined high-level FA categories and enumerated clinical concepts into these categories. Finally, the candidate ontology was sequentially validated over a series of iterations with the panel of experts to include clinical concepts that recorded the highest agreement. The second validation phase involved vetting the dictionary with experts across 22 FCN Centers of Distinction. The data dictionary is organized into 11 categories; Events, Medications, Formal diagnoses, Triggers, Social determinants, Oral Immunotherapy (OIT), Reactions, Procedures, Therapeutic plan, Clinical trial, Unique external data identifier. Each concept has a hierarchical categorization into super classes and subclasses, and concept attributes have been defined to include 977 FA concepts. Although existing clinical vocabularies cover many of these, noticeable gaps are evident, and some concepts are missing entirely, such as “baked eggs” or OIT phases. The FA data dictionary provides a pivotal resource for understanding of complex, multi-institutional data to promote scientific discovery in the field of FA.