There is no single "positive plate" in a module,; although of course the most-positive cell has positive side, it is not isolated any differently from the other cells. Each of the heat transfer fins which runs between a pair of cells and turns to contact the cooling plate is exposed to a different potential - each of them is positive compared to one neighboring fin and negative compared to the neighbor on the other side, if they are not completely isolated from the cells that they are cooling. If the fins where conductively connected to the cell terminals, any conductive plate clamped to them would be a short across every second cell of the module, and across the entire module. If the plates were conductively mounted to the frame, the entire pack voltage would be shorted through the frame.
The outside of each pouch is intended to be isolated from both terminals, but that isolation is not expected to be complete - there is certainly capacitive coupling, but there is also some conductivity.
In this design, which is widely used in LG Chem and other modules, there is no case between the cooling plates and the whatever surrounds the module. The modules are not intended to be exposed, but are supposed to be (as they are in snowdog's car) enclosed in a controlled environment, in contact with a heat-transfer plate using suitable materials (and those materials are an aspect which is being worked on). The "case" of the module is really a stack of retention frames, rather than a housing.
