The main method of prevention is blocks on users that make malicious edits. Obviously that doesn't prevent all malicious edits.

The larger users of the data are doing some degree of validation of changes before showing them to users.