The real answer is to use a filesystem that doesn't require off-line integrity checks, and to disable automatic fsck when you're stuck with something that does.

I use ZFS, which does this sort of thing in the background while the system is live. In fact, every read from a ZFS filesystem does a complete integrity check of everything that read depends on. If it finds an error during that process it fixes it before returning data to the caller, or returns a read error if it cannot. Those errors get collected and logged, so that if they do occur (usually due to faulty hardware) you can recover the effected files from backup.

This is much better than a surprise fsck when you're running late.