Skip to main content

Table 2 Under- and overtriage rates of logistic regression and XGBoost (median and 2.5 and 97.5 percentiles (calculation on 1000 runs))

From: The advanced machine learner XGBoost did not reduce prehospital trauma mistriage compared with logistic regression: a simulation study

Events per free parameter Data set Undertriage logistic regression Overtriage logistic regression Undertriage XGBoost Overtriage XGBoost
10 SweTrau 0.321 [0.259, 0.389] 0.321 [0.299, 0.344] 0.324 [0.258, 0.683] 0.319 [0.057, 0.344]
10 NTDB 0.429 [0.338, 0.79] 0.453 [0.052, 0.501] 0.701 [0.35, 0.808] 0.08 [0.039, 0.494]
25 SweTrau 0.314 [0.257, 0.379] 0.322 [0.3, 0.346] 0.316 [0.258, 0.61] 0.321 [0.09, 0.345]
25 NTDB 0.405 [0.332, 0.771] 0.46 [0.06, 0.499] 0.436 [0.345, 0.792] 0.444 [0.045, 0.498]
100 SweTrau 0.312 [0.254, 0.373] 0.323 [0.301, 0.345] 0.314 [0.255, 0.4] 0.322 [0.291, 0.345]
100 NTDB 0.394 [0.324, 0.735] 0.466 [0.072, 0.503] 0.409 [0.327, 0.79] 0.459 [0.048, 0.497]
1000 NTDB 0.395 [0.327, 0.72] 0.468 [0.078, 0.507] 0.406 [0.328, 0.777] 0.463 [0.05, 0.504]
  1. NTDB, National Trauma Data Bank; SweTrau, Swedish Trauma Registry
Back to article page

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us