Fig. 4
Subgroup analyses of the prediction algorithms. We trained the XGBoost model separately for each subgroup to predict hypoglycemia (A) and hyperglycemia (B). The AUC and 95% CI are presented in a forest plot. The numbers of data points and positive labels in each group are also shown. Hypoglycemia: blood glucose level ≤ 80 mg/dL (4.4 mmol/L); Hyperglycemia: blood glucose level ≥ 180 mg/dL (10 mmol/L). AUC: area under the curve; CI: confidence interval, BMI: body mass index; SOFA score: sequential organ failure assessment score. A: Forest plot of prediction models for hypoglycemia. B: Forest plot of prediction models for hyperglycemia.
