Agreement between parental reports and patient records in food allergies among infants and young children in Finland

Abstract

A third of parents suspect food allergy in their children. Questionnaire-based studies usually overestimate the occurrence of food allergies. The aim of the present study was to validate a study questionnaire by comparing children's use of special diets as reported by parents with patient records at the hospital. A population-based cohort with genetic susceptibility to type 1 diabetes (15% of those screened) was recruited in the Tampere area between 1997 and 2001, and followed for development of food allergy for 3 years. Food allergies and other special diets were queried at the age of 3 years with a structured questionnaire. The hospital records of the children, whose parents had reported an elimination diet of the child, were studied to validate the parental reports of food allergies. The hospital database was also checked for the respective diagnosis codes to estimate underreporting. Altogether, 1122 parents returned the questionnaire at the study center visit when the child was 3 years old. Food allergy was reported by 15.0% of the parents. In 10.6% of the children food allergy had been diagnosed or confirmed at the hospital. Hospital-confirmed food allergy was unreported in 0.9% of the cases. The measure of agreement between reported and hospital-confirmed food allergies, using crosstabulation with Cohen's Kappa, was within 0.71-0.88 for cow's milk allergy, 0.74-0.82 for cereal allergy and 0.66-0.86 for any reported food allergy. We found that the validity of the questionnaire obtaining information on food allergies of infants and young children was good to excellent based on a comparison between parental reports and information obtained from patient records.