Skip to main content

Table 4 Simulation results: Mean bias of log(OR) by PSM-based approaches and data generation features

From: Evaluating the performance of propensity score matching based approaches in individual patient data meta-analysis

   

Treatment Prevalence According to Study

50% Treatment Prevalence

30% Treatment Prevalence

OR

SD c

SD b

Across

Preferential

Random

Within

Across

Preferential

Random

Within

Across

Preferential

Random

Within

 

(σ1j)

(σ0j)

Study

Study

Effect

Study

Study

Study

Effect

Study

Study

Study

Effect

Study

3

0.04

0.7

0.09

0.00

-0.17

-0.03

0.11

0.09

0.01

0.05

0.15

0.06

0.04

0.06

  

1.4

0.23

0.06

0.06

0.03

0.32

0.25

0.10

0.15

0.45

0.22

0.13

0.20

  

2.1

0.46

0.22

0.25

0.17

0.58

0.38

0.17

0.19

0.76

0.23

0.18

0.18

 

0.08

0.7

0.09

-0.01

-0.17

-0.03

0.10

0.08

0.00

0.04

0.14

0.05

0.03

0.05

  

1.4

0.22

0.05

0.05

0.02

0.30

0.24

0.09

0.14

0.44

0.21

0.13

0.19

  

2.1

0.45

0.21

0.24

0.16

0.56

0.37

0.15

0.17

0.74

0.22

0.17

0.16

1.5

0.04

0.7

0.07

-0.03

-0.21

-0.07

0.10

0.04

0.00

0.04

0.15

0.06

0.04

0.06

  

1.4

0.23

0.03

0.00

-0.02

0.31

0.29

0.09

0.12

0.43

0.13

0.13

0.18

  

2.1

0.43

0.14

0.11

0.05

0.58

0.45

0.16

0.18

0.75

0.19

0.18

0.14

 

0.08

0.7

0.07

-0.04

-0.22

-0.07

0.10

0.10

-0.01

0.03

0.14

0.05

0.03

0.05

  

1.4

0.22

0.02

-0.01

-0.02

0.29

0.28

0.07

0.10

0.42

0.19

0.12

0.17

  

2.1

0.42

0.13

0.09

0.03

0.56

0.44

0.14

0.16

0.73

0.17

0.17

0.13

  1. OR indicates pooled odds ratio used for simulation
  2. cSD is the standard deviation, σ1j of study specific random slope, u1j
  3. bSD is the standard deviation, σ0j, of study specific random intercept, u0j
  4. Across-study indicates consideration of Single-level logit model with across-study matching
  5. Prefer-study indicates consideration Single-level logit model with preferential-within study matching
  6. Random-effect indicates consideration of Random-effect logit model with across-study matching
  7. Within-study indicates consideration of Single-level logit model with within-study matching