A Rorschach tragedy

Sorry, there will be a bit of a long lead up to get to the point of this post -- skip down if you like.

When a shocking event like the Batman massacre happens, everybody comes along and tries to grind their own ax on it. (Is that grammatically correct?) No, it didn't happen because of teaching evolution, or violent movies and video games, or demonic possession (yup, a Catholic priest came up with that one), or because all men are scum. There's a legitimate argument to be made about regulation of the sale of weapons and ammunition, but tightening up on background checks wouldn't have helped because the perpetrator didn't have any relevant public record.

What happened is, of course, that he had an onset of psychosis at just the age when it usually happens. Advocates for the mentally ill are universally claiming, and getting reporters to include the statement in their stories, that mentally ill people are not any more likely to commit violent acts. I'm afraid that is not exactly true. The truth is that most people diagnosed with schizophrenia are not in any way dangerous, but some are. Many studies have found that severe mental illness is associated with an increased risk of violent offending.

So, this could have been a story about inadequate mental health services and limited access. Except it isn't, because it turns out the perpetrator was seeing a psychiatrist. Everybody has clammed up about this, for two good reasons. He's obviously going to put on an insanity defense, and there's pretty much only one target that victims can sue, and that's the University of Colorado and the psychiatrist.

So now to the point. It has leaked out that the psychiatrist was concerned enough about the possibility that her patient would commit violence that she initiated the university's internal process for evaluating and addressing such situations. When he withdrew from the university, all that evidently stopped. So, did the psychiatrist screw up? Did she fail to take appropriate action, thereby endangering the public?

I can't, obviously, speak to the specifics of this case. But I can tell you that we can't really predict who is going to commit acts of violence. The linked review, in the new BMJ, tells us that the best systematic screening tools will overpredict -- that is, less than half the people identified as high risk will actually offend. And they are expensive and difficult to implement, usually requiring that the person already be in custody.

Just plan clinical judgment, which was all that was available in the Colorado case, is pretty much useless. Psychiatrists, in Colorado as elsewhere, can get a court to commit a person involuntarily but they are obviously reluctant to do this because most of the time, they'll be depriving a person of liberty without good cause.

Every time a criminal on parole commits a serious crime, we get a huge outcry. This just happened in Connecticut, when a guy on parole murdered a gas station owner for a few bucks. But let's face it -- criminals, until they actually do something like murder a gas station owner, will get out of jail eventually. If we could predict which ones would commit violence, maybe we could deny them parole, but a) we can't and b) it's actually better to release people on parole, where at least they get some supervision and conceivably even services, than just to dump them on the street.

So that's it folks. When people do their best in these situations, it isn't always good enough.