Sometimes, an insulation resistance (IR) testing program is in place, yet a major cable failure still occurs. Because it doesn't seem to work, cutting back or eliminating the program appears to make sense -- that is, until you look deeper.
In these situations, what hasn't happened is consistent testing followed by consistent trending. What makes an IR program work is the ability to trend data over time. A proper trend can't exist unless you do the same tests the same way on the same schedule every time. The only variation in the testing should be temperature adjustment, which also must be done in a consistent manner.
Figure out ahead of time what the testing interval will be for a given type of equipment, and stick to that interval. If you don't, then the program will fail as will your equipment. To determine that interval, you must balance a realistic level of access (and required downtime) against the minimum testing frequency that will allow you to detect insulation deterioration in a timely manner.