The evidence I've seen is the review boards are pretty good at determining whether somebody constitutes an ongoing risk, and if they do, they don't release them. The perpetrator in Lori's offence would be a good example of that.
I think it's also important, though, to be able to release them in a graduated way, which this scheme doesn't contemplate. It's a good idea to figure out what their triggers are to ensure they can follow a medical regime, that they stay on their medications, that they don't cross the path of the victims when they've been asked not to. You can do that better when you have a release with conditions.
This regime seems to only contemplate in the hospital or not. I may be misreading it, but I think you probably need that tri-level of graduated release, particularly if someone is known to pose a particular risk. The real problem with the designation is that the proposed paragraph 672.64(1)(b) component which allows you to make a determination based on a single brutal incident that the person is high risk will not hold up. You will not find adequate evidence to suggest and to be able to predict from a single brutal act that this will constitute a future risk of harm. You would know if it happened again it would be very serious, but what you often find is that the more serious the incident, the less likely it is to recur; the less serious the incident, the more likely it is to recur. It's not a straight line.