The Russian satellite "died" long ago.

It's widely assumed that many satellites that "die" are not in fact dead but instead secretly operational. This is part of the reason so much effort is put into accurately tracking "dead" satellites: we want to know which are ones are not as "dead" as claimed.

It's also desirable not to let the bad guys know just how carefully we track "dead" satellites, especially ones that have been "dead" a decade or more - we don't want to give away just how much activity an undead satellite can get away with before being caught.

I don't believe it's particularly hard to find "close calls" in tracking data even for as many objects as the US tracks. The problem is collecting high-quality tracking data in the first place.

There are a lot of things that affect the orbit very slightly but enough to matter (gravity from Earth, Moon, Sun and even Jupiter, atmospheric drag depending on day/night heating and solar activity/CMEs acting on Earth, etc). I have read that the tracking data is no good in less than a week for many objects before the position must be measured again.

There have been enough "no comments" from SPACECOM that I'm assuming they're not going to say whether they suspected a collision was possible or not - they don't want to give away whether we do or don't put that much effort into watching long-"dead" satellites. Even if the data had said a collision was possible they might not have told Iridium in order to preserve the secret, since the odds of actually colliding are so low.