There is another accessibility trait: .playsSound, that does a similar thing than .startsMediaSession. This one seems more suitable for buttons that have their own sound effects when interacting with them, like it often happens in games.

When interacting with a button with VoiceOver, the accessibility label is repeated to the user. If you are playing some audio, it could be difficult to listen to it properly. To avoid that, you can add the .startsMediaSession accessibility trait.