Cohort construction
Children less than 18 years-old were enrolled, and the day of the blood draw for cholesterol measurements was defined as the index date. Subjects with atopic dermatitis diagnosis codes or prescriptions for treatment of atopic dermatitis before the index date were excluded. Children were also excluded if they had previous recorded diagnoses of contact dermatitis, psoriasis, scabies, ichthyosis, erythroderma, photodermatitis, or seborrheic dermatitis. Subjects were separated into a high TC group and a low TC group according to a blood TC level of 170 mg/dL.21 The propensity model and outcome definitions all operate on data that are converted into the CDM.19Thus, there were matching for age, sex, medications, procedures, and measurements based on the medical records of patients in the high and normal TC groups. The association of TC with the onset of atopic dermatitis was assessed using medical data from Kandong Sacred Heart Hospital. The unmatched cohort had 3,913 children with TC levels of 170 mg/dL or more (high TC group) and 48,812 children with TC levels below 170 mg/dL (normal TC group). After 1:4 propensity score matching, the matched cohort had 1,722 subjects in the high TC group and 6,735 subjects in the normal TC group.