Skip to main content

Table 1 Models’ discriminative performance

From: Predicting 7-day, 30-day and 60-day all-cause unplanned readmission: a case study of a Sydney hospital

 

7-day Readmission

30-day Readmission

60-day Readmission

10-fold CV derivation set average

(95% CI)

validation set

10-fold CV derivation set average

(95% CI)

validation set

10-fold CV derivation set average

(95% CI)

validation set

Gradient Tree Boosting

Logistic Regression (selected variables)

Risk Score: RETURN7

Gradient Tree Boosting

Logistic Regression (selected variables)

Risk Score: RETURN30

Gradient Tree Boosting

Logistic Regression (selected variables)

Risk Score: RETURN60

AUC

0.71 (0.70–0.72)

0.68 (0.67–0.69)

0.71

0.74 (0.73–0.75)

0.72 (0.71–0.73)

0.71

0.76 (0.75–0.76)

0.74 (0.73–0.75)

0.74

SEN (%)

54.7 (52.8–56.6)

71.5 (66.9–76.1)

61.5 (Cut-off = 12)

60.3 (59.0–61.5)

59.0 (56.1–61.9)

52.9 (Cut-off = 12)

63.9 (62.4–65.4)

66.1 (63.0–69.3)

70.0 (Cut-off = 11)

SPE (%)

73.0 (71.9–74.1)

56.5 (51.9–61.0)

69.2 (Cut-off = 12)

73.7 (73.3–74.2)

73.3 (70.7–76.0)

77.4 (Cut-off = 12)

73.4 (72.6–74.2)

69.8 (66.5–73.1)

65.1 (Cut-off = 11)

PPV (%)

7.5 (0.7–0.8)

7.3 (6.7–7.9)

6.9 (Cut-off = 12)

16.0 (15.4–16.5)

16.0 (15.4–16.5)

14.8 (Cut-off = 12)

21.6 (20.9–22.4)

21.4 (20.6–22.2)

18.5 (Cut-off = 11)

  1. AUC Area Under the Receiver operating Curve, SEN Sensitivity, SPE Specificity, PPV Positive Predictive Value, Cut-off Score cut-off for which SEN, SPE and PPV are calculated