You’re Assessing Risk Wrong
There is so much anxiety around assessing risk in psychiatry - risk to self, risk to others, risk from others and how to balance the uncertainties of predicting risk and with no wanting to reinforce unhelpful rhetoric or stigmatising beliefs that mental illness inevitably leads to violence. At the core of it is tolerating and managing uncertainty, and recognising that even robust risk assessment does not equal the much-wanted crystal ball of risk prediction. That means there are occasions that we get it wrong, and the post-hoc scrutiny with the benefit of hindsight can go beyond just learning lessons to unhelpfully trying to attribute blame for not predicting somethat that may have been, at the time, unpredictable.
Colleagues at the University of Oxford have published a very novel qualitative study to examine how risk is assessed in early intervention services, who work with people in the earliest stages of their first psychotic episode. About 1:10 people with early psychosis perpetrate some sort of violent act. The research team interviewed 30 clinicians, patients and carers. They found that most risk assessments were unstructured based on clinician experience, and often excluded care givers owing to concerns about patient confidentiality. A lack of structure means some risk factors are not asked about, some are given a greater weight of importance than others
Importantly, clinician confidence in risk assessment was low, particularly around risk to others rather than risk to self where clinician experience was greater. Low confidence leads to hesitation to ask about it. As with many serious incidents, poor information sharing between agencies is also a problem, which in this case mainly comrpsied of limited access to people’s offending history and the risk assessment is therefore mostly reliant on what the patient discloses themselves.
Several risk prediction tools are under study but none have been tested and proved effective enough to be reliably used in clinical practice. Furthermore, we need to be better at how we can then meanignfully address the dynamic and modifiable risk factors for violence beyond just crisis management and in a way that includes patients and their families and carers.
Even if we can’t accurately predict risk, we can be better at structured risk assessments that take the relevant factors into account and in my experience, professionals need to be upfront about the uncertainties in managing risk. If I can leave you with one takeaway message from this blog, is NOT to try and categories the degree of risk as low, medium or high. It’s a fallacy and at some point, you will come unstuck.
