I attended the third Zero Suicide Alliance Conference on 22 January 2020, which had a theme of Artificial Intelligence and Data in Suicide Prevention. I was asked to give the first plenary Keynote Talk, which was titled: Using Data Science to Understand Suicidality and Self-harm.
I discussed how the risk assessment of suicidality is a time-consuming but inaccurate activity for mental health services internationally.
In the last 50 years a large number of tools have been designed for suicide risk assessment, and tested in a wide variety of populations, but studies show that these tools suffer from low predictive values.
More recently, advances in research fields such as machine learning and natural language processing applied on large datasets have shown promising results for healthcare, and may enable an important shift in advancing precision medicine.
I discussed how data science is being used for identification of suicidality and self-harm, using as examples the work we are doing using the Clinical Record Interactive Search (CRIS) system which was developed for use within the NIHR Maudsley Biomedical Research Centre (BRC) but is now also used in other trusts. In our case it provides authorised researchers with regulated, secure access to anonymised information extracted from the South London and Maudsley NHS Foundation Trust electronic clinical records system and enables us to study much richer data in fine-grained detail.
I provided a perspective on the strengths and weaknesses of such approaches to mental health-related data, and suggested research directions to enable improvement in clinical practice.
If you want to catch up on the social media for the day of the conference you can do so by looking up the hashtag #ZSAconf20 run by Andre Tomlin and Vanessa Garrity.