Adjusted odds ratio (aOR) was adjusted for socioeconomic status, number of siblings, family history of food allergy, and infant ancestry (unless stratified by the factor). Firth penalized likelihood logistic regression was used when sample sizes were too small to use a conventional maximum likelihood logistic regression. Error bars indicate 95% CIs.
aStudy years refer to years of enrollment.
bInteraction between the age of peanut introduction and infant ancestry when included in the logistic regression model and assessed using likelihood ratio testing.