Table 3 Scenario (I): performance of the estimated propensity score (PS) model by simulation setting assuming n = 200
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% | -0.001 | 0.002 | 0.034 | 0.112 |
10% | -0.199 | -0.076 | -0.007 | 0.104 | ||
5% | -1.264 | -0.356 | 0.312 | 1.424 | ||
\({\pi }^{*}\) (with \(X_\textit{r}\)) | 50% | 0.030 | 0.029 | 0.048 | 0.105 | |
10% | -0.101 | 0.006 | 0.069 | 0.187 | ||
5% | -1.087 | -0.218 | 0.447 | 1.593 | ||
MSE | \(\pi\) | 50% | 0.229 | 0.245 | 0.336 | 0.865 |
10% | 1.835 | 0.555 | 0.611 | 1.618 | ||
5% | 22.016 | 7.358 | 7.516 | 25.242 | ||
\({\pi }^{*}\) | 50% | 0.187 | 0.202 | 0.271 | 0.711 | |
10% | 1.636 | 0.493 | 0.605 | 1.702 | ||
5% | 20.447 | 6.941 | 8.031 | 27.171 | ||
Coverage (%) | \(\pi\) | 50% | 93.3 | 93.2 | 92.5 | 90.6 |
10% | 91.8 | 92.5 | 93.2 | 93.9 | ||
5% | 88.4 | 91.2 | 92.1 | 85.8 | ||
\({\pi }^{*}\) | 50% | 94.2 | 94.4 | 93.8 | 93.4 | |
10% | 93.6 | 93.2 | 93.2 | 93.6 | ||
5% | 89.3 | 92.8 | 92.9 | 87.2 | ||
Type I error and power (%) | \(\pi\) | 50% | 6.5 | 35.7 | 88.0 | 96.7 |
10% | 8.1 | 21.5 | 74.9 | 94.7 | ||
5% | 11.4 | 16.7 | 51.9 | 78.2 | ||
\({\pi }^{*}\) | 50% | 5.6 | 41.2 | 92.8 | 98.1 | |
10% | 6.1 | 26.4 | 79.3 | 95.5 | ||
5% | 10.6 | 21.0 | 60.3 | 83.1 | ||