Fig. 2: Use of model adjusting for baseline C-peptide and age reduces variance in outcome measure among placebo/control individuals.
From: A standardized metric to enhance clinical trial design and outcome interpretation in type 1 diabetes
Distribution across trials of baseline variables (a) of age and C-peptide AUC mean, and outcome measures (b) of 1 year C-peptide AUC mean and 1-year QR. Use of QR instead of 1-year C-peptide AUC mean reduces standard deviation (SD; blue) for each trial. Mean and SD are indicated by black lines and dark shaded region. Dashed reference lines in panel A indicate average value in combined placebo participants: 16.4 years and baseline C-peptide of 0.73 nmol/L. Dashed reference line in panel B indicates a QR of zero to more easily decipher between positive (above reference line) and negative (below reference line) outcomes. Box plots show the distribution of data, excluding outliers; whiskers indicate minimum and maximum, box indicates 1st and 3rd quartiles (interquartile range), and median is represented by thin horizontal line within box. Violin plots visualize data distribution.
