Google engineers have developed an AI tool that can accurately predict whether a patient admitted to hospital will die within 24 hours.
It’s a new type of algorithm that consumes available data from various hospitals and makes calculations on its own and informs doctors if it’s already too late.
To make its predictions, the Google software uses data such as patient’s ethnicity, age, gender, previous diagnoses, lab results and vital signs.
But what makes it so powerful is that it includes data previously thought out of reach of machines, such as doctor notes buried in PDFs or scribbled on old charts.
As well as death, AI can also predict unplanned re-admissions within 30 days and probable duration of stay at a hospital. The system is still in its infancy, but Google believes it could someday be used to predict death far longer in advance.
There are pieces of evidence to support this. Recently, a woman with the last stage of breast-cancer was admitted to a hospital. Doctors ran several tests on her and hospital’s computers predicted that she had approximately 91% chances of survival.
However, Google’s AI software predicted that the woman had about 20% death risk. The woman passed away after a few days. Google’s software impressed the experts by accessing over 175K data points which were not available earlier. It scanned through PDFs and notes on very old charts – to come up with its own prediction.
What’s more important is that Google did it far faster than existing technology available in the hospital. Associate professors at Stanford University comment that nearly 80% of the time is used in making the data presentable by today’s technology. However, Google’s approach is different. It avoids it and picks up data intelligently.
The next logical step for Google is to deploy its AI systems into clinics to aid doctors with the diagnosis. Google algorithms would assist the experts with diagnosis and even with the treatment by making a large amount of data in easily consumable format.
Just to put the things in perspective, Google’s AI model averaged 0.86 in predicting whether the patients would stay longer in the hospital vs. the existing 0.76. Google’s accuracy was 0.95 in predicting patient’s mortality within a set time frame.
Apparently, AI is making ways into human lives in ways we couldn’t imagine.