I'll add that the only difficulties that have been encountered during the process of installing an IMS is when the engine has already suffered a failure and the inner race and balls from the existing bearing have come out. At this point, it's very hard to get the remaining outer race out, especially on a dual row bearing IMS. Otherwise, it's very straightforward. Regardless, quite a few IMS's that were previously considered gonners have been saved.
There are a few tricks/tools that an adventurous technician can use to extract the remains of a failed bearing but the question remains - what if the debris from the failed bearing has circulated in the engine? Is there any other hidden damage? Also, if the timing was thrown out of whack, there could be some carnage in the form of bent valves, etc., requiring the heads to come off. At this point, it might be best to do a complete teardown rather than trying to use a retrofit kit 'after the fact' to try to resurrect an otherwise failed engine.
To answer the original poster's question on whether a retrofit kit is a permanent fix - we just don't know. We've redesigned the bearing support and relocated the o-ring to make that component stronger and gone to a ceramic hybrid bearing that in the field is usually rated for up to five times the life of a conventional ball bearing. We've used the same sintered silicon nitride material that is used to make the ball bearings, but for cam followers. Our experience was their phenomenal wear characteristics, as in zero wear on full blown race engines with radical high-ramp rate lobes, typically known for severe wear and short life spans. The only drawback was the price.
Here's some info on those ceramic lifters we had years back (no longer available as material became too expensive and hard to get):
http://lnengineering.com/lifters.htm
But that said, obviously the IMS is a flawed item, or Porsche wouldn't have eliminated it from the new DFI engine in the '09 model after having intermediate shafts in their flat 6's for the last 4+ decades. We're just working within the constraints of what Porsche has given us with regards to the intermediate shafts to make it as best as we can.
We're planning a redesign of the flange to address variances in the reluctor wheel (timing teeth) on the flywheel that sometimes causes an interference fit between the two. This has been very rare with only three flywheels having experienced this problem out of the dozens and dozens of kits that have been successfully installed around the country. I'm also working with Flat 6 Innovations to finalize our own puller, to bring the price down and resolve current supply issues, as the pullers currently come out of Germany and take 4-6 weeks for us to get (and are hard to keep in stock).