The dataset comes from The National Map, so I’m not surprised to hear this. TNM’s coverage of fire stations comes from the archived fire station class in GNIS with continuous updates by The National Map Corps in partnership with state and local GIS departments. (Some states are more actively involved than others.) TNM prefers a standard format for fire stations of the form “Denver Fire Department Station 23”. The fully qualified name could go in alt_name=* or official_name=* if it doesn’t go in name=*.
I’m never quite sure what to use as the canonical name=*. Some fire departments have their own naming or numbering scheme among their own stations (“South Station”) while the dispatching center assigns the stations completely different numbers based on a countywide scheme (“Station 42”). Departments can be inconsistent about signposting the department name as part of the station name or leaving it out. Often they don’t signpost anything other than the engine company number, but the station still has a name and/or number. Or the signposted name is an unwieldy memorial designation that no one calls it in reality.
Hewing closely to a monument sign feels pedantic to me, as it does for any community anchor institution. Ease of wayfinding don’t seem so relevant, since a fire station is primarily an origin, not a destination.
As in other HIFLD discussions, I’d encourage you to check out what the upstream source has on these structures. It’s entirely possible that The National Map has more current data for Mississippi, with corrections that HIFLD missed.
Incidentally, the dataset also pulls in The National Map’s coverage of EMS stations. Are you planning to import those as emergency=ambulance_station too? Do you know of any situations in which OSM has identified a whole building as a fire station, but actually it’s occupied by both a fire station and an EMS station, and how would you reconcile that?