Google is training machines to predict when a patient will die by using electronic health record data, and has created a model that is 95 percent accurate in predicting whether hospital patients will pass away 24 hours after being admitted, researchers wrote in a paper published in the journal Nature.
Google's Medical Brain – the division of the company focusing on the medical field – together with colleagues at UC-San Francisco, Stanford Medicine, and The University of Chicago Medicine, engineered a computer system to render predictions off of medical-records data including patient demographics, previous diagnoses and procedures, vital signs and lab results. The A.I. was also able to sift through notes scribbled on old charts or buried in PDFs.
Google obtained data from 216,221 adult patients with more than 46 billion data points for its study, and its deep learning model was able to predict a range of patient outcomes including the length of a patient's stay and their chances of readmission.
'When patients get admitted to a hospital, they have many questions about what will happen next," said Google research scientist and medical doctor Alvin Rajkomar said in a statement to the Daily Mail.
"'When will I be able to go home? Will I get better? Will I have to come back to the hospital?' Predicting what will happen next is a natural application of machine learning."
Google said its model proved to be 10 percent better than traditional models: It scored 0.86 in predicting whether patients would stay longer in the hospital compared to 0.76 via traditional methods and 0.95 in predicting inpatient mortality compared to 0.86 via traditional methods.
"These models outperformed traditional, clinically used predictive models in all cases," Rajkomar wrote in Nature. "We believe that this approach can be used to create accurate and scalable predictions for a variety of clinical scenarios."
© 2025 Newsmax. All rights reserved.