Skip to main content

Table 2 Goodness-of-fit estimates for fixed and random effects fractional polynomial models for different powers p1 and p2.

From: Network meta-analysis of survival data with fractional polynomials

Power p1

Power p2

Dbar

Dhat

pD

DIC

Fixed effects models

-2

-

904.5

884.6

19.9

924.4

-1

-

916.2

896.3

19.8

936.0

-0.5

-

925.5

905.6

19.9

945.3

0a

-

934.6

915.2

19.3

953.9

0.5

-

942.8

923.5

19.4

962.2

1b

-

948.7

929.2

19.4

968.1

2

-

957.3

937.5

19.9

977.2

3

-

964.5

944.3

20.2

984.7

-2

-2

865.5

835.8

29.8

895.3

-2

-1

857.2

827.0

30.2

887.4

-2

-0.5

850.9

821.2

29.7

880.6

-2

0

844.4

815.2

29.1

873.5

-2

0.5

840.2

810.6

29.7

869.9

-2

1

837.1

807.1

30.1

867.2

-2

2

837.3

807.2

30.1

867.4

-2

3

837.8

808.2

29.6

867.4

-1

-1

848.4

819.8

28.6

877.0

-1

-0.5

849.0

820.1

28.9

877.9

-1

0

840.9

812.6

28.3

869.2

-1

0.5

840.2

811.4

28.8

869.0

-1

1

839.8

809.8

30.0

869.8

-1

2

844.3

814.0

30.2

874.5

-1

3

850.4

820.4

30.0

880.5

-0.5

-0.5

840.5

811.6

28.9

869.4

-0.5

0

842.2

813.6

28.6

870.8

-0.5

0.5

842.4

812.3

30.1

872.5

-0.5

1

843.6

813.9

29.7

873.3

-0.5

2

851.9

822.9

29.1

881.0

-0.5

3

861.4

831.4

30.0

891.5

0

0

845.8

817.2

28.6

874.4

0

0.5

849.8

821.6

28.2

878.0

0

1

854.6

823.4

31.2

885.8

0

2

862.5

833.3

29.1

891.6

0

3

874.2

844.6

29.6

903.9

0.5

0.5

854.3

830.6

23.8

878.1

0.5

1

860.0

831.8

28.2

888.1

0.5

2

876.4

846.7

29.7

906.1

0.5

3

888.3

858.8

29.4

917.7

1

1

871.4

842.0

29.4

900.8

1

2

887.2

858.3

28.9

916.2

1

3

902.3

871.7

30.6

932.9

2

2

907.8

880.0

27.9

935.7

2

3

921.2

892.2

29.0

950.2

3

3

934.6

906.9

27.7

962.4

Random effects models

-2

-

904.4

883.2

21.3

925.7

0a

-

934.4

911.4

23.0

957.4

-2c

1

835.9

803.9

32.0

867.9

-2 d

1

836.1

805.1

31.0

867.1

  1. a Corresponding to Weibull distribution for hazard over time; b Corresponding to Gompertz distribution for hazard over time; c Random effects model with heterogeneity for all three effect parameters; d Random effects model with heterogeneity for d0