How to introduce correlations between random effects?

Can we explore the impact of introducing correlations between random effects in PK/PD modeling on the accuracy of predictions and the understanding of complex drug disposition mechanisms?