Well, here is my wild donkey speculation...but let me first re-post a picture of the bridge before collapse:
And here is an extreme closeup of the rocker plate (that is now laying on the ground) on the pier where the collapse seemed to initiate:
Note how the their is a bend in the drain pipe right by the rocker plate...and it looks as if the joint just above the plate was leaking, as the discoloration around the pipe in the area suggests.
I think that pipe was indirectly spilling water all over that plate, which can be an especially bad thing after a harsh winter, as their would be a lot of salt in that water. After forty years of having water dumped on it, I imagine there was a lot of corrosion on this particular plate, perhaps enough to weaken the bolts holding the plate in place.
I think the repairs contributed to the collapse by creating an asymitrical load on the bridge, due to half the lanes being closed, causing a side, horizontal tension that may not have normaly been present when all lanes are open. Also, the vibrations and banging from the repair process might have triggered fractures in bolts holding the plate on the pier. As long as the tension on the plate is moving in a downward direction, no problem as gravity holds it in place...but if there were some sort of horizontal tension, it could have snapped the already weakened bolts.
Also...the pin connecting the bridge to this plate may also have been corroded...and I think that also contributed to the collapse.
A horizontal tension on that plate from the asymitrical traffic load caused the bolts to snap, the plate shifts position just slightly, but enough to send a big shock wave through the bridge...the trusses then start having to deal with a slight horizontal tension they where never designed for, and start crumbling direcly above that plate, twisting and falling. Once that process starts, the rest of the bridge goes, just like kicking a leg out from under a table, and it falls like a row of dominos.
The bridge was designed quite well to handle normal verticle tensions (up and down)...but couldn't deal too well with a horizontal tension (side to side)...at least not a the point where the bridge connected to the concrete piers.
Hopefully an civil engineer will see this thread, and explain it in a more technical fashion... or correct me if I'm wrong.
But that's my humble opinion.