You've included forensic units. When a prisoner in a regular prison develops a severe mental illness they'll be transferred from prison to hospital. Or when someone is found guilty at trial they may be sent direct to a secure mental health unit. Obviously the population of people put in prison includes violent offenders.
You're also focussing on hospitals. Most people with a mental illness, even a severe illness, will not need a hospital admission.
For example: My county has a population of about 650,000 people. The local MH trust has about 4,000 to 5,000 people on their books at any time. Only about 4% of those people need an in-patient stay.
No! And that's the point. Mental illness contributes a tiny amount to violence. Drug or alcohol addiction contribute far more, as does previous violence.
This simply isn't true. There is loads of research on it. People only quote the stuff that reads counter-intuitively, but upon deeper reading into the subject, views like the ones you are expressing are easily de-bunked.
The first link you posted did not aay what you thought it said - indeed it directly contradicted you.
The second link you posted takes almost a third of its data from forensic units.
> Rates of violence would also be expected to be influenced by the type of psychiatric service patients were recruited from. In particular, patients treated in forensic settings are likely to be more violent than those from other settings, not least because the majority are admitted specifically because of their violent behaviour [Coid et al., 2001].
Thus, you include people who are detained against their will by the courts because they pose a risk of harm to others but exclude anyone else with a mental illness and wonder why the rates of violence are so high.
But it's worse than that!
Look at what they count as a violent act:
> The definition of violence and aggression differed widely between studies. Types of violence recorded included physical violence, physical violence directed at staff only, verbal aggression, aggression towards objects, self-harm and sexual aggression.
You're trying to include "self harm" as "violent to other people".
Really, read the article. At least, read the studies before you cite them.
The research is very clear: people with a mental illness are more likely to be the victim of violent crime than the perpetrator; people with a mental illness are far more likely to be violent toward themselves than to other people; mental illness alone provides very little indication on risk of violence, unlike substance-dependance or previous violence.
EDIT: and again this article contradicts your point. Look at the graph on page 12 - most violence is commited by people in a forensic unit. But especially see this quote:
> 3.7 Repeated violence The studies commonly reported that a small sub-group of patients were responsible for the majority incidents, so even with this most-cherry picked sample (people who commit a crime and are then detained in a mental health hospital by the courts) we still see that violence is perpetrated by a small sub-group.
You're also focussing on hospitals. Most people with a mental illness, even a severe illness, will not need a hospital admission.
For example: My county has a population of about 650,000 people. The local MH trust has about 4,000 to 5,000 people on their books at any time. Only about 4% of those people need an in-patient stay.