Assessing dose response from flexible-dose clinical trials is problematic. The true dose effect may be obscured and even reversed in observed data because dose is related to both previous and subsequent outcomes. To remove selection bias, we propose marginal s... ...

请注册登录后继续浏览