Errors in medical literature: not a question of impact