Deep Learning Algorithm to Suitably Predict Palliative Care Needs

Deep Learning Algorithm to Suitably Predict Palliative Care Needs

Palliative Care

Scientists from the Stanford University have devised an AI algorithm that can, based on a number of factors including response to drugs, family history, and the nature of the illness, predict how long can people live. This study can help patients and caregivers understand better when and how much palliative care is important. The paper, revolving around the concept of improving palliative care for patients with the help of deep learning.

It is estimated that nearly 60% of patients spend their final days at acute care hospitals, which provide aggressive care to patients, trying to make the last few days or months of a critical patient as free as possible from suffering. Acute-care healthcare providers work towards managing factors such as loss of appetite, pain, nausea, confusion, while also supporting the patients morally and psychologically, respecting their and their families’ spiritual, cultural, and social needs.

Researchers state that the algorithm has not been devised as a way of replacing doctors towards the end of patients’ lives but with the view of providing a way of improving access to palliative care. It is in a way an opportunity for both the healthcare providers and patients to have necessary end-of-life conversations while there is still time. Understanding the correct time to choose palliative care is necessary for the fact that if the patients are admitted well before time, it can place an unnecessary burden on the healthcare systems and monetary burden on patients and their families.

If the patients are admitted too late, on the other hand, they miss out on this stage of care. The neural networks that were used to build this Deep learning algorithm made use of previous years’ electronic health record (EHR) data.