Statistical Analyses
Statistical analyses were performed with R software (2020, version
4.0.0, R Core Team, Vienna, Austria). Data on patient demographics,
comorbidities, and management were presented with proportions for
categorical data and medians with an interquartile range (IQR) for
continuous data. Subgroup analyses were performed for patients
presenting with only CSU and patients presenting with CIndU (regardless
of co-existing CSU). To ensure the validity of the UCT among children in
whom the CDLQI is validated, a subgroup analysis was performed in
patients aged 4 to 16 years.
Internal consistency of the UCT was evaluated by Cronbach’s α
coefficient, calculated by the Psych package. Cronbach’s α
coefficient was interpreted as follows: <0.60 unacceptable,
0.60-065 not desirable, 0.65-0.70 minimally acceptable, 0.70-0.80
respectable, 0.80-0.90 excellent, and >0.90 excessive
redundancy17.
Convergent validity of the UCT was assessed by Pearson partial
correlation with the CDLQI using the ppcor package in groups with
≥25 patients18. Partial correlations were controlled
for sex and age. The following strata were used to interpret r values:
0.90-1.00 very strong correlation, 0.70-0.89 strong correlation,
0.40-0.69 moderate correlation, 0.10-0.39 weak correlation, and
0.00-0.10 negligible correlation19.
Known-groups validity of the UCT was assessed by the Kruskal-Wallis test
using established CDLQI disease severity strata: 0-1 no effect, 2-6
small effect, 7-12 moderate effect, 13-18 very large effect, and 19-30
extremely large effect20.
We considered a CDQLI cut-off value of ≥6 as indicative of poorly
controlled disease21. Using this cut-off, we assessed
the screening accuracy and the area under the curve (AUC) of the UCT in
the identification of poorly controlled CU by receiver operating
characteristics (ROC) curve via the pROC package. Optimal UCT
cut-off values were estimated by the Youden index22.
AUC values were interpreted according to the following cut-off values:
≥0.90 outstanding, 0.80-0.89 excellent, 0.70-0.79 acceptable, 0.50-0.69
poor, and 0.50 no better than chance23.
Bootstrapping hypothesis testing with 100, 000 iterations was used to
evaluate the change in UCT scores from study entry to follow-up between
patients who started or increased their dose of sgAHs versus patients
who decreased their dose or ceased sgAHs.