library cache miss ratio very high - how to find out the rot cause

Hi we are driving a Documentum based ECM system on a big sun machine (12 CPU) (i know that Documentum is creating very bad SQL) Right now sometimes (quite often) the system is totally slow. With some monitoring we can see that at that point the miss ratio for the sql library cache goes over 60 %. At that time the shared SQL area of the SGA is used only by 60 % and