Would you like to know the exact day of your death? Scientists at Stanford have been working on an AI program that can do that. Although it can sound a bit off-putting to the average person, the researchers see it as a positive development that can help people with terminal illness to find closure.
The team at Stanford created an algorithm which was fed the health records of roughly 2 million patients from two hospitals. The algorithm learned to estimate mortality with the following three to 12 months. Research scientist and lead author of the study, Ken Jung, told Newsweek that their goal is to help better inform palliative care specialists. This will help more patients make important decisions that impact how their death will be handled and prepare their families for life without them.
“All too often, advanced illness turns to a medical crisis, and patients end up in the ICU. There, events can attain a momentum of their own, resulting in increasingly aggressive interventions that do not serve patients and their families well,” he said.
As IEE Spectrum reports, normally palliative care specialists wait until a patient is recommended to them by a primary care medical team. But that recommendation is normally delayed in favor of aggressive medical procedures designed to prolong their lives. However, the algorithm flips the process around so that the palliative care team can proactively contact a patient or their family when they are nearing an end of life situation.
According to Spectrum, the project was the brainchild of Stephanie Harman, an internal medicine physician and founding medical director of Palliative Care Services for Stanford Health Care. She and her colleague Andrew Ng, an adjunct professor and former head of the Baidu AI Group, both decided that they would explore ways that artificial intelligence could be applied to palliative care.
The algorithm uses deep learning, a machine learning process that utilizes neural networks to sift and gain knowledge from huge amounts of data. A neural network is a programming paradigm inspired by the interconnected neurons in the brain. The team says that the large scope of the data they collected allowed them to predict the time of death from a wide range of conditions. It was not focused on a particular disease or demographic.
The full text of the research can be found in a journal article named, “Improving Palliative Care with Deep Learning” which can be found on the arXiv pre-print server.