AI could be used to identify the Huddersfield patients at risk of suicide

Artificial Intelligence
Image credit: source

Advances in artificial intelligence could be the key to identifying people at risk of suicide.

Technological developments represented by AI are at the heart of an assessment tool, soon to be trialled in Kirklees, that would use GP practice data to automatically compute the suicide risk factor for patients.

It would enable doctors to get warnings of patients most at risk of suicide who need referral or an increased focus.

The work, described as a “high risk decision support tool”, is being spearheaded by University of Huddersfield academic Professor Grigoris Antoniou.

A globally-renowned expert in AI technology, Dr Antoniou has led the way in pioneering work into predicting which mental health patients are most likely to take their own lives.

Rebecca Elliott, public health manager for Kirklees Council , said the authority was investigating how to focus on primary care to improve suicide prevention.

Speaking to members of the Health and Adult Social Care Scrutiny Panel (Jan 21) she said: “Primary care is one of the areas that we want to try and improve our suicide prevention approach to.



Prof Grigoris Antoniou, a globally-acknowledged expert in AI technologie at the University of Huddersfield,s, who has led the pioneering use of artificial intelligence in predicting which mental health patients are most likely to take their own lives

“We know from previous suicide audits that a significant number of people will attend primary care before they go on to take their own life.

“What they go to the GP to talk about, we don’t know, but we can tell when we look at our suicide audits that there may be frequent visits that take place. “

The tool has already been trialled in secondary care “with good results” and Professor Antoniou is now looking for the support of a handful of GP practices across Kirklees to follow suit.

Added Ms Elliott: “It’s not to replace the GP’s role at all. It’s about making it easier to find who on their records may be at risk of suicide. The way that his model is put together is based on all sorts of demographic factors.

“It doesn’t mean that that person is going to take their own life. It just means that they have significant risk factors for suicide.

“So we’re trying to help the GP find who that person might be to try and facilitate and inform a conversation to bring the subject up.”

Mike Doyle of South West Yorkshire Partnership NHS Foundation Trust said Dr Antoniou’s work was “very promising”.

He added: “There is still some way to go.

“The key to it is learning the lessons from people that have taken their own life.

“That’ll then help clinicians and GPs, in this case, make decisions about future care and treatment.”

You don’t have to suffer in silence if you’re struggling with your mental health. Here are some groups you can contact when you need help.

Samaritans: Phone 116 123, 24 hours a day, or email jo@samaritans.org, in confidence

Childline: Phone 0800 1111. Calls are free and won’t show up on your bill

Platform 1 men’s community group: Support for issues including mental health problems and addiction recovery. Visit the website or call 01484 421143.

Andy’s Man Club: info@andysmanclub.co.uk

PAPYRUS: A voluntary organisation supporting suicidal teens and young adults. Phone 0800 068 4141

Mind: A charity offering support and advice for people with mental health problems.

Students Against Depression: A website for students who are depressed, have low mood, or are suicidal. Click here to visit

Bullying UK: A website for both children and adults affected by bullying. Click here

Campaign Against Living Miserably (CALM): For young men who are feeling unhappy. There’s a website and a helpline: 0800 58 58 58

MindOut: Provide support and advice on mental health for members of LGBTQ communities. Phone 01273 234839

Follow ExaminerLive

(Excerpt) Read more Here | 2020-01-22 20:16:00

Leave a Reply

Your email address will not be published. Required fields are marked *