What can be achieved with a small standard clinical dataset?