Unrealistic expectation of risk prediction by mental health services

The tragedy of the Nottingham homicides and other attempted murders by Valdo Calocane has highlighted the role of the police and mental health services in preventing such incidents (see eg. Times article). The police have admitted they could have done more to locate him as he was wanted for arrest after assaulting a police officer, but have argued that even if he had been arrested that it was highly unlikely he would have received a custodial sentence (see another Times article). The police also said that he would probably not have engaged with mental health services if re-referred to them, as he had not done so previously.

Calocane has now been detained in a High Security hospital (see judge’s sentencing remarks). It seems that he has had a chronic psychotic illness which has been difficult to treat. Many people with a psychotic illness are lacking insight into their illness and, as they do not see themselves as ill, do not engage well with treatment. Very few of these seriously mentally ill people are dangerous, though, like Calocane.

Assertive outreach (AO) was introduced as part of the NHS plan from 2000 to manage these difficult to engage patients through providing intensive home treatment by specialist teams. Many NHS Trusts have now closed such services as it has been difficult to show in research studies that such specialist teams reduce bed use or produce better clinical outcomes compared to standard services. Nonetheless, research has found that AO generally improves the engagement of such severely mentally ill patients and their satisfaction with services. The public does expect services to keep in touch with the most severely mentally ill, so I think it has been a mistake to close AO services. Too many people have been discharged from mental health services, when they shouldn’t have been, for not engaging with services. 

More generally, people do need to have a more realistic attitude to what mental health services can do to prevent such serious incidents. It’s commonly thought that all that needs to happen is that mentally ill people need to take their medication and everything will be alright. But even if they take their medication, not everyone gets better.

Homicide inquiries in mental health services have often been destructive (see my unpublished paper based on 2013 conference presentation). They all tend to find the same kind of things, for example: poor risk management, communication problems, inadequate care planning, lack of interagency working. The trouble is that these sort of things are everyday difficulties that staff have to manage. Some evidence of the need to improve in those areas will often be found in all cases, not only the ones that lead to homicide. Communication, care planning and joint working can always be improved. Usually this situation does not lead to adverse outcomes.

Moreover, homicide inquiries have encouraged the myth of accurate risk assessment. Mental health services are expected to have foresight which it is unrealistic to expect them to have. Not uncommonly homicide inquiries, because of hindsight bias, unreasonably recommend improved risk assessment. But, merely knowing about risks, even if these could be calculated, may not actually improve practice at all. Such recommendations also encourage an inappropriate separation of risk assessment from an overall full and complete assessment, which is, therefore, not completed properly because the focus is on mere risk assessment. The reality is that we have very little knowledge about the accurate quantification of risk. Risk assessment is actually the management of uncertainty, not certainty. Mental health services should have and still need to challenge the illusion of certainty.

The media are often quick to blame mental health services for homicides by psychiatric patients (eg. see previous post). I do think it is reasonable that someone like Valdo Calocane was sent to a special hospital rather than prison. Mental dysfunction can diminish responsibility for a crime (see eg. another previous post). What people often do not appreciate is that people detained in a special hospital may spend longer there than if they had gone to prison. Their mental state may not improve sufficiently for them to be released and I suspect this may well happen with Calocane.

Mass shootings, particularly in USA, although people run amok in similar ways in all societies (see previous post), are often blamed on poor mental health services (see previous post). Actually the majority of mass shooters are not psychotic, like Calocane. Generally they are aggrieved individuals who blame others for their suffering, as well as themselves, and commonly mass shooters at least attempt suicide after their killing of others. Too often mental health services inappropriately accept the blame for such outrages, apparently with the misguided hope that doing so will increase the social and political will to improve mental health services (see another previous post). Psychiatry does have a role in preventing homicide and suicide but it does need to be far more realistic with the public about what it can achieve.