I would think you could merely have the booster turn on and off --- still keeping a "straight" line, but changing the predicted arc from what was predicted early on (which admittedly would be longer than doing it right the first time) (looks like it would hit Chicago, but goes for Houston).
OR, have the re-entry warhead come down at a steeper-than-free-fall angle by the reverse --- e.g., a missle looking like it would hit Houston, hit Chicago (comes over the NP).
Missle defense will need to be improved so that it can deal with ICBM's that can change course. It may not be possible to hit them but it may be possible to force them to change course to such an extent that they miss the intended target. Also, if a warhead coming toward the U.S., it is very doubtful that we would rely on a single hit-to-kill unit or even a single method to try to take it out. We'd probably fire off everything we had. If this is done in the correct manner, at the various stages, then it become even more difficult for the warhead to evade and still be able to hit the intended target.
The comet impact mission a few months ago used similar technology. The warhead has rockets on it to move it side to side as well as forward and back during the coast phase, and enough fuel to give it some crossrange capability. This happens after the booster is done and the warhead has separated and is on its own. The ABM interceptors also use this technology. Whether an ABM impactor can hit the warhead depends on which has more fuel and stronger rockets.
I guess it boils down to what the Russians think that "we" think when a misslie looks like its headed someplace predictable...