ACC	accuracy
SEN	sensitivity
SPE	specificity
TP	number of true positives, i.e., patient presenting severe complication correctly predicted as positive
TN	number of true negatives, i.e., patient without severe complication correctly predicted as negative
FP	number of false positives, i.e., patient without severe complication wrongly predicted as positive
FN	number of false negatives, i.e., patient presenting severe complication wrongly predicted as negative
Total	total number of the patients, i.e., TP + TN + FP + FN
P	number of patients presenting severe complication, i.e., TP + FN
N	number of patients without severe complication, i.e., TN + FP
AUC	area under the receiver operating characteristic (ROC) curve for binary outcome
T	threshold for a patient is classified as presenting severe complication if X > T, where X is predicted probability of a patients presenting severe complication by an algorithm.