The team of computer scientists and medical experts from Google and three universities in the United States (US) reportedly found a way to predict the death of patients in the hospital.
Software developed that relies on artificial intelligence technology or AI. Researchers claim AI-based software with ‘in-depth learning’ capabilities can predict patient outcomes over other methods currently available.
“This model outperforms traditional predictive models in all cases,” wrote a statement from the newspaper.
To make this prediction, software uses medical record data including patient demographics, previous diagnosis and procedures, lab results and vital signs. The top part of the predicted outcome list is “inpatient mortality,” in which the patient is reported as “expiration”.
But sometimes the software goes beyond life and death questions, to answer important questions for patients and hospital administrators. The machine also reportedly predicts the return of patients to medical facilities within 30 days, possible stay and diagnosis and hospital billing codes.
The research paper studied 216,000 inpatients and 114,000 patients unknown to researchers at two hospitals owned by UC San Francisco and the University of Chicago.
“The biggest claim is the ability to predict patient deaths 24-48 hours before the current method, which allows time for doctors to manage life-saving procedures,” according to Quartz online magazine, which saw the paper published on January 26.
The software is claimed to be able to predict deaths with an accuracy of 93 percent to 95 percent, about 10 percent better than traditional prediction methods, according to the newspaper.