Skip to main content

Table 1 Comparison of the performance of different estimation methods of the effect of cumulative exposure to psychosocial stressors at work on blood pressure in the parametric plasmode simulation scenario

From: Longitudinal plasmode algorithms to evaluate statistical methods in realistic scenarios: an illustration applied to occupational epidemiology

 

Intermittent

Cessation

Onset

Chronic

Bias

    IPTW

-0.37

-0.03

0.23

-0.05

    g-computation

0.27

0.02

0.16

-0.01

    TMLE

0.26

-0.09

0.24

-0.07

    TMLE+SL

0.25

-0.09

0.24

-0.07

    Baseline adjustment

-0.40

-0.11

0.48

0.01

    Baseline and Time-varying covariate adjustment

-0.44

-0.08

0.21

-0.19

Standard deviation

    IPTW

1.29

1.05

1.27

1.29

    g-computation

0.53

0.84

0.86

1.05

    TMLE

1.20

1.00

1.05

1.18

    TMLE+SL

1.20

1.00

1.05

1.18

    Baseline adjustment

1.12

0.98

1.06

1.17

    Baseline and Time-varying covariate adjustment

1.11

0.98

1.06

1.19

RMSE

    IPTW

1.35

1.05

1.29

1.29

    g-computation

0.59

0.84

0.88

1.05

    TMLE

1.22

1.00

1.08

1.18

    TMLE+SL

1.22

1.00

1.08

1.18

    Baseline adjustment

1.19

0.98

1.16

1.17

    Baseline and Time-varying covariate adjustment

1.20

0.98

1.08

1.21

Coverage of the 95% confidence interval

    IPTW

93%

96%

95%

95%

    g-computation

90%

94%

94%

94%

    TMLE

92%

94%

93%

93%

    TMLE+SL

92%

94%

93%

93%

    Baseline adjustment

92%

95%

92%

95%

    Baseline and Time-varying covariate adjustment

92%

95%

94%

94%

  1. Legend: IPTW = Inverse probability of treatment weighting, RMSE = root mean squared error, SL = Super Learner, TMLE = Targeted maximum likelihood