Multilayer well screen simulation

Hello,
I've got a question about simulating well screens with Multilayer Well Boundary Condition.

I need to simulate pumping and reinjection wells at different depths inside a multilayered model domain and I'm actually assigning the MW boundary condition only to layers that effectively correspond to the well screen (5 to 20 meters long, starting from the bottom of the well) and not to the entire well bore depth (sometimes 5 to 10 times longer).
This way the upper vertical, containing the rest of the (non screened) well casing, is not considered at all in the flow simulation.

Is this the best way to simulate multilayer wells with short well screens?
Should I try instead to consider also the well casing and the disturbed zone around the well or the effects will be negligible in medium to large scale models?

Thank you!

Another related (and interesting) issue would be how to simulate the effective pumping depth along the well screen. The pump is not always placed in the lowest node of the simulated vertical...

Unless you look at very local effects, it is sufficient to consider the screened interval of the well. As the conductivity in the open pipe is very high compared to the aquifer, the head difference within the well pipe will be very small. thus the vertical location of the pump (FEFLOW assumes the lowest node) is typically not crucial. If for some reasons you think it might be, you can do a comparison by using a Well BC plus discrete features for the well instead of a multilayer well. Technically this is identical, but you can place the Well BC in any node you want.

Thank you Peter,
I'll try to make some tests, maybe exaggerating the well screen length and placing the pump at the top of the screened interval; I'd like to verify the results in term of contaminant plume interception at different depths.
It would be very nice to have the same advanced level of simulation (and interface) available in Feflow for BHE (well construction details, materials, filter pack, geometry and so on) also for standard extraction and injection wells; for example, it should help to verify the real efficiency of a P&T system in a small scale model.