We build a covariate model to find the effect of any covariates on a pk parameter. Many times it has been observed that covariates has statistically significant effect while modelling, but in real life the covariate effects were modest (≤ 20%) and not clinically meaningful. Why so?
Can anyone explain this?