When the women with advanced breast cancer reached the city hospital, her lungs were already flooded with fluids. After a radiology scan and seeing two doctors, it was estimated that there is a 9.3 percent chance of her dying during her stay. A new algorithm created by Google was put to work and after a thorough assessment, she has 19.9 percent risk of dying. She died within a few days.
This shocking medical case was published in May by Google. It highlights the healthcare possibilities held by neural networks. It is a kind of artificial intelligence that uses a host of data to learn, develop, and predict. This tool of Google can predict a number of patient outcomes including their duration of stay in the hospitals, chances of readmission, and risk of dying.
The fact that was most intriguing to the medical experts was the ability of Google to go through even the inaccessible data like notes in PDFs or old charts. The neural net used all the scattered information to provide verdicts with high accuracy and minimum time taken. The system showed the documents and factors it has used to reach the conclusions. For years, the medical experts have been trying to get hold of comprehensive electronic data of patients to offer better treatment options. It has been seen that the availability of more data in time not just saves lives but also enables the doctors to concentrate on patients and not on paperwork.
The next step for Google is to bring in the system of predictions into the clinics. The current work is made possible by Google’s health research unit, also known as the Medical Brain. It is working on a number of AI tools that can predict diseases and symptoms with such a high level of accuracy that it is alarming. Ever since Google declared its focus on AI, it has been trying to improve existing services on the internet. The advancements brought forth by Medical Brain has let Google venture into a new market which is something the co-founders, Larry Page and Sergey Brin had always wanted to achieve.