Skip to main content

Table 1 Scenario (I): performance of the estimated propensity score (PS) model

From: Treatment effect estimation using the propensity score in clinical trials with historical control

Performance measurement

PS model

Outcome event rate

Odds ratio

1.0

2.0

5.0

10.0

Bias

\(\pi\) (without \(X_\textit{r}\))

50%

0.004

-0.016

-0.034

-0.032

10%

-0.026

-0.022

-0.018

-0.013

5%

-0.060

-0.031

-0.009

0.017

\({\pi }^{*}\) (with \(X_\textit{r}\))

50%

0.035

0.014

-0.005

-0.007

10%

0.019

0.020

0.023

0.030

5%

0.000

0.023

0.045

0.074

MSE

\(\pi\)

50%

0.045

0.050

0.069

0.097

10%

0.117

0.092

0.086

0.097

5%

0.228

0.175

0.160

0.201

\({\pi }^{*}\)

50%

0.037

0.039

0.052

0.074

10%

0.097

0.080

0.084

0.100

5%

0.187

0.154

0.160

0.217

Coverage (%)

\(\pi\)

50%

95.0

94.6

94.0

93.2

10%

93.9

94.3

94.8

94.7

5%

93.2

93.9

94.3

94.4

\({\pi }^{*}\)

50%

94.9

94.8

94.8

94.3

10%

94.7

94.9

94.6

94.4

5%

94.3

94.1

94.5

94.1

Type I error and power (%)

\(\pi\)

50%

5.0

86.6

99.8

100.0

10%

6.1

63.7

100.0

100.0

5%

6.8

41.7

98.5

100.0

\({\pi }^{*}\)

50%

5.0

94.1

99.9

100.0

10%

5.2

72.4

99.8

100.0

5%

5.7

50.7

98.4

99.8

  1. \(\pi\) (without \(X_\textit{r}\)): the conventional method; \({\pi }^{*}\) (with \(X_\textit{r}\)): the proposed method