Covariate adjustment of treatment effects in clinical trials

摘要: Abstract It has been previously shown that the magnitude of adjustment a treatment effect Z value for baseline covariate depends on both degree disparity between groups and correlation outcome variable. In this article, is further explored tabulated variety situations. Theoretically, may be great, example, from = 3.0 unadjusted to 13.1 adjusted, or 2.0 − 1.6 adjusted. However, in more practically realistic situations will generally fall range 1 2 standard error units, when as great ± 3. If there perfect comparability with respect covariate, still sizable difference adjusted effect, almost always greater absolute than value.