For every doctor, regardless of their expertise or long-time practice, one of the hardest tasks will always be to give professional judgment of the patient that has no chance to survive. More difficult than that is telling the truth to his family.
But what if, instead of a doctor, this decision can be made by a machine in the future? What if artificial intelligence could take the burden from a family that needs to sign a consent and unplug the device?
One innovative Beijing hospital introduced an AI system that made, in at least seven cases, diagnoses different from the ones made by the most expert doctors, and it was right.
The Chinese Academy of Sciences, in collaboration with PLA General Hospital, has worked hard for over eight years to create Artificial Intelligence that will have the ability to make accurate predictions about patients in a coma.
Its accuracy in prognostic assessments was a staggering 90%, after being tested in the real environment of Beijing hospital.
Best neurologists evaluated patients who experienced severe brain trauma and were in a state of coma. Their potential recovery was recorded on a scale from 1 to 23.
In the case of a young man who fell into a vegetative state after a serious accident, doctors gave seven points, signifying almost no chance of awakening and recovery. However, the machine rated his brain scan by 20 points and was right.
Also, at least six more cases were reported at the same hospital, with equally severe head injuries and no chance of recovery for which the computer had predicted a happier end that eventually happened.
However, during the testing, the AI machine also gave a misjudgment. It had given the low number of points to one patient who fully recovered within a year.
This system is different from the one used to detect cancer, for example. Magnetic resonance scanning can capture even the smallest changes in brain functions that are not visible to the human eye. The AI algorithm records all neurological activity and finds matches between spotted changes and previously determined patterns.
Therefore, machine learning can embrace a greater range of information and find, with more precision, the possibility of waking up from coma and trauma recovery.
Machine learning can significantly change the image of clinical treatment, not only in China but around the world as well. However, an intelligent machine is not always capable of understanding and processing all the facts to make an accurate assessment. In some cases, patients die under unplanned circumstances, like sudden allergy or pneumonia. In others, AI may lack the essential data and medical history, especially with patients that are coming from less developed areas.
AI cannot yet assume the full responsibility of deciding about a patient’s future, but in collaboration with the best and most experienced specialists, it can save many lives.