A woman with late-stage breast cancer came to a hospital, saw two doctors and got a radiology scan. The hospital’s computers read her vital signs and estimated a 9.3 percent chance she would die during her stay.
Then a Google algorithm read up on the woman — using 175,639 data points — and assessed her death risk: 19.9 percent. She died within days.
The account was published by Google highlighting the health care potential of neural networks, a form of artificial intelligence software that uses data to learn. Google had created a tool that could forecast a host of patient outcomes, including how length of hospital stays, odds of readmission and chances they will soon die.
What impressed medical experts most was Google’s ability to sift through data previously out of reach: notes buried in PDFs or scribbled on old charts. Google’s next step is moving this predictive system into clinics, said AI chief Jeff Dean, whose research unit is referred to as Medical Brain.
C-sections not tied to overweight kids
A study in JAMA Pediatrics found there was no significant difference in obesity at age 5 between babies born vaginally and those born by cesarean section. It used a database to study 16,140 siblings born between 1987 and 2003 and their 8,070 mothers. Several studies had suggested that babies born by C-section are at higher risk for obesity, perhaps because of differences in the babies’ microbiomes.