Death – as inevitable as it is – never fails to make us curious about when our time might come. But as it turns out, artificial intelligence might have some answers for you.
Stanford researchers have been able to develop an AI that can predict the time a patient has left, with up to 90 percent accuracy. As unnerving as the idea might sound, the team behind it claim how it could vastly improve end-of-life care for patients and their families too.
Also read: Humans and AI having babies
This is primarily to help caregivers prioritize the wishes of terminally or seriously ill patients as the AI has already accurately pinpointed the time they might pass. It can also allow them to ensure that important conversations can be held before it's too late.
In the study's pre-print to arXiv, the team explained how the last days of terminally ill patients are often lived in a way that is in striking contrast to how they would actually like to live them out. And as per researchers, roughly 80 percent of Americans want to spend their final days at home.
Yet almost 60 percent of them end up dying at the hospital. In an effort to close this gap, the particular AI has been designed.
"We could build a predictive model using routinely collected operational data in the healthcare setting, as opposed to a carefully designed experimental study," Anand Avati, a PhD candidate in computer science at the AI Lab of Stanford University, told IEEE.
"The scale of data available allowed us to build an all-cause mortality prediction model, instead of being disease or demographic specific."
The tool doesn't work by itself to guide the care process; instead, it is used together with the assessment made by the doctors and the combines analyses in making decisions about the end of life planning.
It is also not easy to understand who needs palliative care and when, as the researchers explained. "The criteria for deciding which patients benefit from palliative care can be hard to state explicitly,' the authors explain in the paper.
"Our approach uses deep learning to screen patients admitted to the hospital to identify those who are most likely to have palliative care needs. The algorithm addresses a proxy problem – to predict the mortality of a given patient within the next 12 months – and use that prediction for making recommendations for palliative care referral."
Yet the system, even though useful, also has a challenge that it needs to overcome. Based on the 'black box' nature of the algorithm, the researchers aren't aware exactly what it bases its predictions on.
But why it made the predictions is not necessarily important either. "The palliative care intervention is not tied to why somebody is getting sick,' research scientist Kenneth Jung told IEEE. If it was a different hypothetical case of 'somebody is going to die and we need to pick treatment options,' in that case we do want to understand the causes because of the treatment."
He adds, "In this setting, it doesn't matter as much as long as we get it right."
