Read on Twitter
This week's #tweetorialsinpsychiatry is on suicide risk assessment in the ED (or indeed anywhere) and is predominantly aimed at non-psychiatrists and psychiatrists in training. I hope it is of help. #psychtwitter #medtwitter #medstudenttwitter #suicide #depression #psychosis
For additional reading around this tweetorial I recommend the following guide from the Oxford Centre for Suicide Research: http://cebmh.warne.ox.ac.uk/csr/clinicalguide/docs/Assessment-of-suicide-risk--clinical-guide.pdf
and
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4937078/pdf/40596_2015_Article_434.pdf
and
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4937078/pdf/40596_2015_Article_434.pdf
There are ~6000 suicides/yr in the UK (~47k in the US) of which ~90% have a psychiatric disorder. Depression accounts for 60%, Half have seen their GP within 3 months and 20% in the week before death.
3/4 of all suicides occur in men. Suicide is the leading cause of death in males <35. The most common method in the UK is hanging followed by self-poisoning. In the US firearms account for about half. (Data from the above suicide risk assessment guide)
The UK five year forward view for Mental Health (MH) called for the DH, PHSE and NHSE to support areas to have multi-agency plans in place to reduce suicide by 10% by 2020/21.
The table below outlines some important risk factors from the aforementioned Oxford risk guide. Table reproduced with permission from Professor Keith Hawton at the Centre for Suicide Research, Dept of Psychiatry, Oxford.
Historically, there have been three approaches to assessing suicide risk: 1) unassisted clinician prediction, 2) use of standardised scales/biological tests, and 3) scales derived from statistical modelling.
Clinician prediction into traditional categories of low, medium and high risk is poor. A previous National Confidential Inquiry into Suicide found that over 80% of patients who ended their lives were felt to be at 'no or low' immediate risk at their last review.
Trying to predict risk using standardised rating scales is just as problematic. This is because the positive and negative predictive values of these scales are dependent on the prevalence of the outcome and suicide is a rare event.
A recent systematic review and meta-analysis of PPV's for risk scales demonstrated a pooled PPV estimate of 4-21%. The false +ve and false -ve rates of these scales are too high for clinical use.
DOI: https://doi.org/10.1192/bjp.bp.116.182717
DOI: https://doi.org/10.1192/bjp.bp.116.182717
It is for this reason that the National Institute for Health and Care Excellence (NICE) currently recommends that:
Because of this there has been a move away from risk prediction towards risk formulation. One model for doing this aims to condense elements of the history into 4 distinct domains: Risk status, risk state, available resources, & foreseeable changes. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4937078/
Risk status includes 4 domains: past suicidal acts, impulsivity incl. substance use, long term risk factors, strengths/protective factors. These are judged relative to others in a specific sub-population.
Risk state refers to more dynamic factors including: recent suicidal ideation/behaviour, stressors, precipitants, symptoms, suffering & recent changes, engagement & alliance. These are judged relative to the patients baseline.
Available resources: these are elements available to the patient and mental health team to support treatment planning and safety eg family & friends, social services, housing, financial advice, church groups, chidlcare/respite etc
Foreseeable changes include events/stressors that could reasonably be expected to increase or decrease risk e.g., anniversary of significant events, relationship breakdown, exam results, housing/benefits review, legal proceedings.
An introductory video to this approach can be found here: https://vimeo.com/105130731
The most common scenarios in which suicidal thoughts might arise are in patients with depression, personality disorder and psychosis. Suicidal thougths should always be enquired about. Asking about suicidal thoughts does not increase the risk of suicidal acts.
It is easier to ask about these in a room where you will not be disturbed and after having establised a reasonable rapport with the patient. Use a form of words that feels right for that particular patient at that point in time.
I would usually ask something like 'I was wondering if this is all getting a bit much at the moment?', followed by 'do you sometimes have thoughts about ending your life?', and 'have you thought about how you might do it?'
'do you think it is likely that you might harm yourself in the near future?', 'what is stopping you from doing it at the moment?' 'what do you think might make it hard for you to resists harming yourself over the next few days/weeks?'
This tweetorial topic has been particularly stubborn in its refusal to allow me to reduce it to 3000 characters yet still do it justice. Above I have covered the theoretical elements but the most important thing is to understand how these ideas might be applied in practice.
To that end I recommend you access via the folowing link to my One Drive account where I give non-psychiatrists a glimpse under the bonnet of how we tend to formulate risk and management plans in practice. https://1drv.ms/w/s!AjnRAIR_Os7KgkrtSdjDEGTd01jc
Finally, thank you for reading another tweetorial and I hope it has been as helpful to you in reading it as it has been to me in preparing i.
The second from last tweet with the One drive attachment is the key part of the tweetorial - just in case it gets skipped!
