Skip to main content

Table 2 Goodness of fit over the entire time horizon

From: Comparing current and emerging practice models for the extrapolation of survival data: a simulation study and case-study

Overall mean

Sample size: 100

Sample size: 300

Sample size: 600

squared error

FU: 2 years

FU: 3 years

FU: 4 years

FU: 2 years

FU: 3 years

FU: 4 years

FU: 2 years

FU: 3 years

FU: 4 years

Damped trend, local level

0.51

0.34

0.42

0.23

0.38

0.29

0.26

0.41

0.27

Current practice

1.01

1.19

1.26

0.94

1.15

1.19

0.90

1.12

1.15

Royston-Parmar model

1.98

2.38

1.87

2.21

2.38

1.50

2.25

2.36

1.40

Damped trend, global level

3.75

4.98

2.36

7.88

2.29

0.52

8.07

1.41

0.35

Local trend, local level

3.33

4.41

2.96

6.86

4.13

1.26

9.18

3.39

0.71

Local trend, global level

6.03

7.12

4.27

15.61

6.67

1.36

18.04

4.65

0.57

Generalised additive model

32.89

18.16

6.85

18.49

6.59

2.12

20.27

4.09

1.53

Fractional polynomial: order 1

312.40

103.82

22.49

326.43

41.25

8.61

331.78

35.71

9.14

Fractional polynomial: order 2

531.90

258.30

147.35

205.23

55.05

85.21

121.62

24.07

65.57

Overall bias

 Damped trend, local level

0.38

−0.03

− 0.19

− 0.12

− 0.35

− 0.40

− 0.30

−0.30

− 0.31

 Current practice

−0.36

− 0.37

− 0.35

− 0.55

−0.55

− 0.54

−0.60

− 0.58

−0.56

 Royston-Parmar model

−0.92

−1.07

− 1.10

− 1.07

− 1.10

− 1.11

−0.88

− 0.79

−0.77

 Damped trend, global level

−0.35

− 1.83

− 1.80

− 1.17

−0.78

− 0.56

−0.76

− 0.18

−0.14

 Local trend, local level

−0.93

− 1.80

−2.14

− 1.32

−1.36

− 1.23

−1.06

− 0.64

− 0.48

 Local trend, global level

− 1.36

− 2.85

−3.13

−1.73

− 1.72

−1.30

− 1.31

− 0.45

− 0.11

 Generalised additive model

−1.55

−2.18

− 1.99

− 0.06

− 0.09

0.05

0.15

0.31

0.23

 Fractional polynomial: order 1

−10.52

−11.87

− 12.24

−5.36

−4.10

− 3.91

−2.37

−1.75

− 1.83

 Fractional polynomial: order 2

−5.45

−8.03

−6.72

1.45

− 0.79

−1.84

4.05

3.53

3.08

  1. FU Follow-up