Machine learning to decipher suicidal tendencies in patients

Technology - Akanksha Singh - Nov 10,2016

Rating -

stars - based on 21 reviews

machine learning helps in deciphering suicidal tendencies

It’s been quite some time since technology has invaded every-single part of our lives.  In today’s time, we can’t even step outside the house without our smart phones.

But it’s still a wonder when we hear that robots can now perform heart surgery; so, it is only a medical miracle, when one hears that a ‘computer’ can decipher if a person has suicidal tendencies. A study, published in the journal ‘Suicide and Life Threatening Behavior’ showed that machine learning can be up-to 93% accurate in the classification of a suicidal person.

Taking its note from written or spoken words, machine learning technology can be very helpful for physicians to make out if a person has suicidal tendencies or not. The clinical trials lead author John Pestian said, that the results provide evidence strong enough, to start using advanced technology as a decision-support tool for helping physicians and caregivers identify and prevent suicidal behavior.

John and his team enrolled 379 patients from emergency departments, inpatients and outpatients centers at three sites. The enrolled patients belonged to three different categories, ‘suicidal’, ‘mentally ill and not suicidal’, and ‘neither’. Each patient went through a complete and standard ‘behavior rating scale’, participated in a semi-structured interview, where questions like, “Do you have hope?”, “Are you angry?” and “Does it hurt emotionally?” were put forward

The team of researchers then extricated and analyzed verbal and non-verbal language through the same. Final stage analysis involved the use of ‘machine learning’ algorithms to decipher the results and classify the patients into the three categories. The results proved that machine learning is up-to 93% accurate, in knowing the difference between these groups; great day for technology and medicine to jump into bed together.