[Sapientoni]

I retired as an industrial bearing predictive maintenance specialist. I had equipment to analyze and predict bearing failures so they could be scheduled for replacement and preventing production interruptions of a $900 an hour employee expense. Depending on the mechanics doing the replacement, this could take from 1 hour to 7 hours. I was the only mechanic who used the 1 hour time method. I got a little dirty, but everybody got dirty doing this repair.
The factory 987.1 is the only IMSB I have experience with. Bearings fail because of "spalling" of the surfaces that are loaded by the operation being supported. Tensioning of the bearing in the 987.1 by the chains puts loading on basically 180° of the stationary inner race. I have observed personally the failure of the 987.1 bearing at 148,000 miles due to the spalling of this surface. Due to the factory design, my suggestion is to turn the stub shaft, without removing the flange, at any clutch change. Usually this might take place around 70-90,000 miles? Turning the stub presents a "fresh", previously unloaded surface for the balls to roll on. This is preventive "tuning" of the bearing.
FYI, the factory IMSB for the 987.1 has 8 balls instead of the usual 7 balls. This earns it the "high capacity" designation. It's a few dollars more when purchasing. My observation of the ceramic bearings available are that they all contain 7 balls. RE: ceramics. Very hard, but possibly very brittle? Correct installation handling is imperative to get what life they will give when installing.