Covariance and variance evaluations of two estimators for drug–placebo difference in a trial with sequential parallel design |
| |
Authors: | Yanning Liu |
| |
Institution: | 1. Janssen China, Xinmei Union Square, Shanghai, China;2. Jassen Pharmaceutical L.L.C., Titusville, USA |
| |
Abstract: | Chen et al. Contemp. Clin. Trials, 32: 592–604 (2011)] heuristically proved that the covariance of two estimators is zero assuming equal correlation coefficients. In this article, the above covariance is analytically and rigorously re-derived without any strong assumption in equality between two correlation coefficients. Under rigorous analytic derivations plus assuming number of subjects continuing into Period 2 is a random variable, covariance is reconfirmed to be zero for both normal and binomial data. |
| |
Keywords: | Drug–placebo difference placebo effect seemly unrelated regression sequential parallel design |
|
|