Qb Rating is really flawed, it's a tool, but shouldn't be the thing you point to and go "see, he's (declining, improving, bad, etc etc)". If you don't believe, just look at Brady's QB rating in the Superbowl.
In fact, his 95 QB rating in the Superbowl surprised me so much that I played around with the rating calculator. Believe it or not, but Brady could have thrown for 8 TDs and 0 INTs, with the rest of his stat line the same, and STILL had a lower QB rating than Ryan... It's junk.
QB rating doesn't take into account decision making, such as taking horrible sacks, not throwing the ball away when you should, handing off the ball for a rushing TD or even rushing it in yourself. I agree that team offense needs to be greatly factored into evaluating a QB because at the end of the day, not only is that the ONLY thing that matters, but a team does what is in the best interest of winning, NOT piling up stats. If running the ball makes more sense situationally, Brady will hand it off.
Finally, efficiency is great, but it doesn't give all that much information. For instance, you could have a game where it's heavy run game with a bit of play action, so you end up with a super efficient game and scoring wasn't really a problem. QB rating would like stellar. But then you play a game against a good defense, where you can't run the ball, at all. And they take away the middle of the field and deep stuff. So you have to dink and dunk and work to get the yards and score. You may throw twice as many passing TDs, but because of the game plan, that was necessary, and led to points and a win, you get heavily 'penalized' in the stats category with efficiency.
QB rating is ONLY good to see if a QB was overall good or bad over a large enough sample size. It is completely useless to gain anything more precise than just a 'squint'.