Some good practices I'm aware of are:
As you mentioned, maneuvers are simulated before they are commanded and their effect is evaluated on ground so that thruster parameters and tank filling are updated, so if anything funny is happening during maneuvers this can be identified.
If propulsion is electric (which is still not so common), then thrusting is almost continuous, and generally autonomous with ground supervision.
Geostationary spacecraft are usually in a "collocation window" where they may be close to other geostationary satellites. The basic technique to ensure two satellites are far from colliding is the eccentricity-inclination vector separation.
Operators need to communicate so each one can know their agreed eccentricity and inclination vectors (though I dislike this nomenclature), and larger or smaller distances between spacecraft need to be considered in some cases due to RF interference.
Sometimes, a dead satellite or debris may approach the orbit of a satellite cluster. The operators need to inform each other on the maneuvers they'll be doing, and that there is an agressor object approaching, just in case one of them didn't heard about it.
Maneuvers with liquid propulsion are scheduled to take place on fixed days of the week. This way, the operators know that Tuesday is longitude correction day, while Friday is wheel unloading day. Routine is good for these people.
Also, when working with liquid propulsion, maneuvers are usually of three types: North-South, East-West and wheel unloading.
Hope this helps.
EDIT: Reaction wheel accumulate angular momentum over time, but their maximum speed is limited. Because of this, and due to the lack of significant magnetic fields in geostationary orbits, it is necessary to use propulsion to provide a torque, which, when compensated by the reaction wheels, allows them to have their angular momentum reduced, and thus within their speed limits. This is called "wheel unloading" because the loaded momentum is removed from the wheels.